Research ArticleDecade-Long Changes in Disparity andDistribution of Transit Opportunity in Shenzhen ChinaA Transportation Equity Perspective

Qingfeng Zhou 1 Donghui Dai 1 Yaowu Wang 1 and Jianshuang Fan 2

1Harbin Institute of Technology Shenzhen Graduate School Shenzhen Guangdong 518055 China2Zhejiang University of Technology Hangzhou Zhejiang 310014 China

Correspondence should be addressed to Donghui Dai dai donghuihotmailcom

Received 2 April 2018 Revised 26 July 2018 Accepted 7 August 2018 Published 2 September 2018

Academic Editor Paola Pellegrini

Copyright copy 2018 Qingfeng Zhou et alThis is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Efficiency and equity have always been the two points of focus of transport projects Compared with efficiency equity is easilyoverlooked in the evaluation of transport projects Many studies emphasize that defining and operationalizing costs and benefitsand the distributive principle are critical parts in the assessment of transportation equity However the scope and time frameof the assessment target are also critical In this paper we took Chinarsquos fastest urbanizing city Shenzhen as a case study toassess transport equity by comparing accessibility among groups First the public transport system was divided into bus andsubway and the residents were divided into two groups urban village and nonurban village Second we adopted an enhancedpotential opportunity model to measure residentsrsquo bus and subway accessibility and summarized them as transit opportunityThird we used the Dagum Gini coefficient decomposition and kernel density estimation method to explore the fair distributionof transit opportunity among groups and districts from 2011 to 2020 Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture On the one hand the development of Shenzhen public transport system had a positive effectAll populations are benefiting and their accessibility is increasing On the other hand it also had a negative effect to exacerbateinequality between populations For the absolute value of the opportunity Shenzhenrsquos urban village populations do have fewertransportation opportunities than nonurban villages and this gap between themwill be wider more andmoreThe public transportsystem is more inclined to improve the population with high initial opportunity and make them higher The results illustrated theimportance of examining transportation equity over an extended period and could provide information on urban developmentstrategies

1 Introduction

Public transport is an effective way to solve the problem oftraffic congestion and environmental pollution in high pop-ulation density metropolitan More importantly it providesthe necessary motorized transport to access jobs and socialactivity needed especially for low-income people withoutcars [1] Many cities are aware of the importance of urbanpublic transport and are planning to enhance public transportservices but the improvement of transit may not help low-income people Decision-makers need to know the scope andscale of the benefit of people from the public transportationsystem to make more sensible investment decisions for an

equitable transportation system In recent years the public isincreasingly concerned about the impact of traffic policy andinvestment on equity

Transport-related equity involves a wide range of topicsand previous studies can be divided into four areas (1)research on the consequences of transport inequity andpeople who are vulnerable to suffer transport inequityexploring the relationship between transit supply and time-poverty social exclusion and well-being [2ndash5] (2) study onthe conceptual frameworks to integrate equity assessment intransport project appraisals such as discussing cost-benefitanalysis (CBA) and multicriteria analysis (MCA) which ismore suitable for transport equity assessment [6ndash8] (3) focus

HindawiJournal of Advanced TransportationVolume 2018 Article ID 7127342 16 pageshttpsdoiorg10115520187127342

2 Journal of Advanced Transportation

on the match between transit supply and transit demandbased on spatial mismatch theory concerning access todifferent opportunities among socioeconomic population [9ndash11] (4) research on equity aspects of distribution effects ofpublic transport policies and infrastructure projects throughaccessibility [12ndash14] guiding assessing the distribution ofbenefits generated by transport investment projects for trans-port agencies In this field many studies recognized thatdefining and operationalizing costs and benefits and thedistributive principle are critical parts in the assessment oftransport equity [15] However the scope and time frameof the assessment target are also critical At present thetopic of incorporating equity consideration is involved verylittle in transport projects evaluation and decision-making inChinaThis paper focuses on the fourth area taking the Chinafastest urbanizing city Shenzhen as a case studyThe purposeof this paper is to assess transport equity from changes ofdisparity and distribution of transit opportunity over theperiod from 2011 to 2020 Does the development of transithave a different impact on different group of people and towhat extent What is the trend of change in transit equityeffect

This paper is organized as follows Section 2 presents aliterature review including equity types variables and mea-sures involved in transport equity analysis Section 3 brieflydescribes the variables and measures used in this paper anenhanced transit opportunity measure the decompositionof the Gini coefficient and the kernel density estimationmethod Section 4 introduces the basic situation of publictransportation in Shenzhen the data used in the research andthe concept of ldquourban villagerdquo to classify people In Section 5we assess transport equity from changes of disparity anddistribution of transit opportunity in Shenzhen FinallySection 6 summarizes the results of this study and providesdirection for improving the equity of transit opportunitydistribution in future studies

2 Previous Work

21 Type of Transportation Equity Conducting transportequity analysis first involves conceptual issues of equity Thedefinition of equity has extensive discussions in all fieldsfromphilosophy to economics It differs in different historicalperiods and different perspectives of research The definitionof equity used in this study is ldquothe distribution of benefitsand costs over members of societyrdquo [16] There are twomajor types of transport equity horizontal equity and verticalequity [17] Horizontal equity advocates that individuals orgroups are considered with the same weight consideredequal in ability and need transport should provide serviceequally regardless of need or ability and avoid favoringspecific individual or group over another Vertical equityrecognizes that the ability and needs of individuals or groupsare not the same transport should favor spatially econom-ically and socially disadvantaged individual or groups tocompensate for overall inequities Vertical equity comprisesthree components spatial social and economic Spatialequity refers to providing the equitable transport services and

improvements in transport infrastructure especially inperipheral or rural areas Economic equity is related totransport services designed for users with lower incomesor without transport affordability Social equity is relatedto the availability of special transport services adapted forpersons with mobility impairments Vertical equity can bedivided into equity of opportunity and equity of outcomefrom another dimension Equity of opportunity means thatdisadvantaged people have adequate access to social activityopportunities Equity of outcome means that society mustensure disadvantaged people succeed in social activitiesMany works of literature agree that people should haveequity of opportunity scholars focus on equal opportunityrather than equitable outcome this study also holds thisview

22 Analysis of Transportation Equity Three issues need tobe clear when conducting an equity assessment of trans-port policy or infrastructure projects [18] Frist it is aboutinequality of what variable We need to define costs andbenefits of transport system second it is about what unitof inequality measured The definition of equity introducedbefore involves the categorizing people we need to choosetarget population groups over which costs and benefitsare distributed third it is about the distributive principleused to judge what distribution of costs and benefits isldquomorally properrdquo and ldquosocially acceptablerdquo The distributionprinciple is related to the type of equity and equity mea-sures

221 The Variable of Costs and Benefits in TransportationEquity Many variables can be used to represent the costsand benefits in transportation equity and the most oftenused are transport affordability access to transport andaccessibility to opportunities Transport affordability mea-sures individuals or householdrsquos actual expenditure on publictransport usually as the percent of household disposableincome [19] Public transport affordability equity addressesthat a city should enable their poorest citizens to afford atleast the motorized trip rates reached by average incomestratum Equity strongly relates to distribution effects so themain limitation of affordability is that it does not reflectthe distribution effect well Access to transport is the abilityof a person to reach transit facilities [20] It measurestransport service characteristics and physical proximity totransport service Many papers use it to assess the equityof public transit service provision [21ndash23] Its disadvantageis that it does not consider whether the transport systemcan provide the individual with the desired destinationHaving transport services does not mean that the transportsystem can enable people to go to the destination wherethey want to go to Accessibility is the ability and easeof achieving activities opportunities and goods which isfrequently used to evaluate transit provision concerningequity [14 24 25] It can reflect the distribution effectMartens K [19] claims that accessibility is the most appro-priate measure of benefits from transportation plans andinvestments

Journal of Advanced Transportation 3

Accessibility to opportunities is related to cumulative-opportunity and potentialgravity measures which sum thenumber of destinationsjobs reachable within certain timesby transport mode substantial literature discusses measure[26] This measure is particularly useful in describing howwell transportation networks perform about the distributionof destinations and the needs for subgroups Measure acces-sibility to opportunities usually requires an attractivenessindicator a transport network and an impedance function oftravel costThe attractiveness indicator is expressed regardingthe number of jobs or variables that represents opportunitysizeThe transport network is used to obtain travel costs andtravel cost can be travel time distance fares or a combinationof the three [27 28] There are many ways to get travelcosts such as simulation of the transport network in trafficsoftware calculation from simplified transport network inGeographic Information System (GIS) or calculation froma realistic transport network information database Theimpedance function is an inverse function which indicatesthe increase in travel costs will reduce the opportunities ofattraction

222 The Categorizing People in Transportation EquityEquity analysis needs to define a unit that can be distin-guished and units usually are groups of peoplehouseholds orregions Many studies use demographic and geographic fac-tors categorizing people to identify transport disadvantagedpeople in equity evaluation [17] and these factors includeincome car ownership age gender career household com-position and location of residence Most papers distinguishbetween high income and low income as well as car ownersand car-less individuals or households Indeed disadvan-taged status ismultidimensional some studies combine thesefactors to determine whether a person or area is a transportdisadvantaged

223 The Equity Measures in Transportation Equity Well-known horizontal equity measures are the Gini coefficientTheil index and coefficient of variation they are expressed asratios which are compared among groups to measure equityperformanceGini coefficient initially indicates level of equal-ity of income distributions in economic studies it rangesfrom0 to 1 0means absolute equality and 1 indicates absoluteinequality in transportation equity it is used to evaluate thedegree of accessibility concentration level of different regionsor groups of people and compare the level of equity beforeand after implementing a policy or transport infrastructureSome argue that the Gini coefficient fails to indicate thestructure of inequality the same Gini coefficients can havedifferent income distributions by the group Theil index isusing the information entropy concept to measure individualor interregional income inequality named The Theil indexhas good decomposability as a measure of inequality whenthe sample is divided into multiple groups The Theil indexcan measure the contribution of the intragroup gap andthe intergroup gap to the total gap so it provides moreinterpretation of the inequality among different groups Theprimary approach of vertical equity is to evaluate transport

policy or infrastructure projects according to how they affectaccessibility between disadvantaged peoplehouseholds orregions It is fairer if transportation disadvantaged groupbenefits like transport service improvements favor lower-income areas and groups or transportation services providemore access to job opportunities and other ldquobasicrdquo activi-ties

The keys to equity analysis of public transportationare the measure of accessibility and the equity measureof distribution Our work complements previous researchfrom four aspects First when evaluating the equity impactof public transport only one mode of public transportis concerned in previous studies so the result of equityevaluation could be bias We combined subways and busesto consider equity issues in this paper and proposed anenhanced potential opportunity model to measure residentsrsquobus and subway accessibility considering public servicereliability attractiveness and frequency Second due to thelimitations of data acquisition scholars discuss the impacton equity of transport policy or infrastructure projectsduring a relatively short period (before and after the imple-mentation of the target) Our research used a long-termdata to examine equity situation dynamic change and itwas helpful to capture the trends of equity influence ondifferent groups in the development of public transportThird indicators of transport distribution effects were furtherexplored and applied We used the Dagum Gini coefficientdecomposition which is more convenient thanTheil measureand kernel density estimation method to investigate thefair distribution of potential opportunities At last existingresearch focused on Europe and the United States and wetook the China fastest urbanizing city Shenzhen as a casestudy to assess transport equity by comparing accessibilityamong social groups The results can provide a referencefor the study of the impact of transportation equity in theworld

3 Methodology

In this study the public transport system was including busand subway and we divided the residents into two groupsurban village population and nonurban village populationwhich will be explained in Section 4 We used an enhancedpotential opportunity model to measure residentsrsquo bus andsubway accessibility and summarized them as transit oppor-tunity Then we used the Dagum Gini coefficient decompo-sition and kernel density estimation method to explore thefair distribution of transit opportunity among groups anddistricts from 2011 to 2020

31 Measure of Accessibility This research adopted cumu-lative-opportunity and potentialgravity measures modelsto measure transit-based job accessibility and made someenhancements Given that accessibility measurement isespecially important for the analysis of traffic equity thissection will detail how this research calculates accessibil-ity

4 Journal of Advanced Transportation

Step 1 Calculate the service range of each transit stop Theservice radius of the bus stop is 500 meters and the serviceradius of the metro station is 700 meters

Step 2 Calculate the population in each transit stop servicearea and calculate the job opportunity in each transit stopservice area In our study job opportunity is represented bythe floor area of factory company and government office

Step 3 For transit line 119897 consider both 119894 and 119895 are transit stopsof line 119897 and calculate 119860 119894119895119897 the job opportunity of 119894 can beassessed at 119895 as follows

Calculate per capita service frequency

119878119894119895119897 = 119881119894119895119897119880119875119894 (1)

119881119894119895119897 is average of vehicles (one day or a week) of transitline l from 119894 to 119895 U is transit vehicle capacity and 119875119894 is thepopulation of transit stop 119894 119878119894119895119897 reflects different transit servicecapabilities of bus and subway

Calculate 119879119894119895119897 travel time from 119894 to 119895119879119894119895119897 = 119879119886119888119888119890119904119904 + 119879119908119886119894119905 + 119879119894119899minusV119890ℎ119894119888119897119890 + 119879119890119892119903119890119904119904 (2)

The access and egress times are assumed to be 5 minof walking time which transit users are generally willing toundertake waiting time at transit stops is assumed to be one-half of the scheduled headway when the average headway oftransit service is around 10min and the in-vehicle travel timeis calculated using the scheduled arrival and departure timethat is obtained from transit service schedules

Calculate the distance decay factor

119891119894119895119897 = 11 + 120572119890minus120573119879119894119895119897 (3)

120572 and 120573 are two coefficients which need to be calibratedand we used an average travel time survey The survey datashow that in Shenzhen 963 of the people make their worktrips in less than 60 min Therefore we assumed that theconnectivity from an origin to a destination that takes 60mintravel time would be 0037 The estimates of the parametersare 120572 = 00024321 and 120573 = -0143161

Calculate the job opportunity of 119894 which can be assessedat 119895

119860 119894119895119897 = 119878119894119895119897119874119895119891119894119895119897 (4)

where Oj is the opportunity at 119895Step 4 Consider the job opportunity of 119894 can be assessed at 119890that requires a transfer between transit lines 119897 and119898

119860119896119894119890 = 12 (

119860 119894119896119897119891119894119896119897 +

119860119896119890119898119891119896119890119898 )119891

119896119894119890 (5)

119891119896119894119890 = 11 + 120572119890minus120573(119879119894119896119897+20+119879119896119890119898) (6)

where 119894 is the transit stop of line 119897 119890 is the transit stop ofline119898 and 119896 is the transfer center between 119897 and 119898

Step 5 Calculate the cumulative opportunities of i

119860 119894 = sum119895

119860 119894119895119897 +sum119896

sum119890

119860119896119894119890 (7)

Step 6 Convert residential area to the centroid and calculatethe sum of job opportunities of bus stops in the 500-meterbuffer of residential centroid and the number of metro jobopportunities of metro stations in the 700-meter buffer ofresidential centroid (if there are more than 1 transit stopsbelonging to the same line in the buffer they will be averagedas 1 transit stop) The sum of bus and metro opportunities isthe transit opportunity for residential areas

32 Decomposition of the Gini Coefficient The method ofdecomposition of the Gini coefficient in a discrete space isproposed by Dagum [29] The Dagum Gini coefficient iscalculated using the following

119866 = sum119896119895=1sum119896ℎ=2sum119899119895119894=1 sum119899ℎ119903=1 10038161003816100381610038161003816119910119895119894 minus 119910ℎ1199031003816100381610038161003816100381621198992119910 (8)

119866 is the Gini coefficient 119910119895119894(119910ℎ119903) is individual income of119894(119903) belong to subgroup 119895(ℎ) 119910 is the average income of allpopulation 119899 is the number of all population k is the numberof subgroups and 119899119895( nh) is the number of people in thej(h) subgroup Dagum decomposes the Gini coefficient intothree components [26] (1) 119866119908 contribution ofwithin groupsincome inequalities to 119866 (2) 119866119887 the net contribution of thebetween-group inequalities to119866measured on all population(3) 119866119905 the contribution of the transvariation between thesubpopulations to 119866 Detailed derivation and calculationprocess for each component are listed in [29] the equation ofGini coefficient decomposition in three components is shownas follows

119866 = 119866119908 + 119866119887 + 119866119905 (9)

Decomposition of the Gini coefficient not only effectivelysolves the source of group disparities but also describesthe distribution of subgroups and solves the problem ofoverlap between groups (shows the structure of inequality)In this paper the Gini coefficient of transit opportunity iscalculated and decomposed the population in Shenzhen isdivided into different groups it helps us to know if the transitopportunity gaps within groups generate the inequalities orif the transit opportunity gaps between groups engender theinequalities

33 Kernel Density Estimation Kernel density estimation(KDE) is a nonparametric way to estimate the probabilitydensity function of a random variable in statistics [30] Let(x1 x2 x119899) be a univariate independent and identicallydistributed sample drawn from some distribution with anunknown density Its kernel density estimator is shown asfollows

119891ℎ (119909) = 1119899ℎ119899

sum119894=1

119870(119909 minus 119909119894ℎ ) (10)

Journal of Advanced Transportation 5

where K is the kernel a nonnegative function thatintegrates to one the normal kernel is often used h gt 0 isa smoothing parameter called the bandwidth

As mentioned in Section 22 Gini coefficient or Theilmeasure are ratios which are compared among groups tomeasure equity KDE enables us to shape the distribution of avariable and analyze differences from the perspective of visualgraphics comparison So we use KDE as a complementaryway to draw probability distribution curves of transit oppor-tunity of different population groups and grasp disparity anddistribution of transit opportunity changes in decade-longurban developments in Shenzhen

4 Study Area

41 The Social Group Shenzhen is located in the Pearl RiverDelta region with a land area of 19968 km2 and an urbanpopulation of over 14 million in 2016 It is the first SpecialEconomic Zone (SEZ) city after the institution of reform andthe Open-Door Policy in China in 1979 In the past 30 yearsthe operation of a market economy has made Shenzhenrsquoseconomy develop rapidly bringing with it a dramatic popula-tion increase and spatial expansion In the study of transportequity an important part is to group residents accordingto their socioeconomic level In Shenzhen detailed data onresidentsrsquo occupations and income is not easily accessible sowe use three characteristics of a residentrsquos residence to reflecthisher socioeconomic status These three characteristics areaverage house price of residence the average rental price ofresidence and whether the residence is in the urban villagethe third feature especially is an essential basis for judging aresident as disadvantaged people

Urban village (Cheng Zhong Cun in Chinese) somescholars preferring the term ldquourbanized villagesrdquo or ldquovil-lages in the cityrdquo to avoid the confusion with the West-ern planning concept ldquourban villagerdquo is an outcome ofChinarsquos rapid urbanization and its associated rural-urbanmigration When urban expansion encroaches into ruralland the city government needs to acquire land rightsfrom the rural collective to convert the rural land intourban land The city government only expropriates thefarmland of the village to avoid the costly compensationto relocate villagers and the housing land remains in thehands of the collective Over time the village settlement issurrounded by urban built-up area creating the so-calledurban village The bond between the urban village and thedisadvantaged people is mainly caused by the residenceregistration (hukou) system and rural-urban migrationtide in China The hukou system divides the populationinto the rural population and the urban population Changefrom the rural to urban population needs to be approvedby the authorities Rural migrants can stay in large cities astemporary residents (do not have a local urban hukou) andthey are excluded from the formal urban housing marketCommercial housing is generally expensive and thus notaffordable to migrants who are employed in low-paid jobsMore affordable units provided by urban housing provisionsystem generally require a local urban hukou and are thus

not available to rural migrants Chinarsquos land policies haveenabled the native farmers in the urban villages to constructinexpensive housing units and rent out these units to the ruralmigrants In many cities the urban village is a significanttype of settlement for both local landless peasants andmigrants which are two groups with high urban povertyincidence

The data supplied by the Shenzhen Urban PlanningBureau (SUPB) and the Urban Planning and Design Instituteof Shenzhen (UPDIS) shows that there are 2942 urban villageresidential lands and 4683 nonurban village residential landsin Shenzhen 2009 The Municipal Building Survey 2009provides information for all buildings in Shenzhen includingthe urban villages There are 615702 buildings in 2009 and333576 (54) in urban villages Urban villages in Shenzhenwhich are thought to accommodate approximately sevenmillion meet the basic needs of people particularly poor andlow-income residents [31 32] There are also some studiesproving that immigrants in urban villages are vulnerableConcerning economic sectors the largest proportion ofmigrants is employed in retail hotel catering and otherservices (508)The second most common category is man-ufacturing (193) and construction (92) The proportionof people employed by highly paid public and finance sectorsis tiny [32] With the relatively poor employment profileincome among migrants is low in comparison with the cityaverage In 2004 the Municipal Government found that theaveragemonthly income amongmigrants was only 1149 yuanfar below the average personal income in the city (2195 yuan)(Shenzhen Municipal Government Housing System ReformOffice 2004)

Our study first divided the population into two groupsresidents living in urban villages called the urban villagegroup (UVG) and residents not living in urban villages calledthe Not urban village group (NUV) The UVG is mainlycomposed of low-income migrants and contains some high-income local residents A study shows that the ratio oflocal residents to migrants in the urban village is 188 [32]Thereforemost of the residents in the urban village group arelow-income people and they are more dependent on transitto access jobs and social activity The residents in NUV live informal urban houses meaning that they have owned a houseor can afford the higher rental price (the rent of a formalhouse is 25-5 times the rent in the urban village in Shenzhen)so they have higher disposable income than UVG Secondaccording to housing prices and rents in different districtsof Shenzhen each group is further divided into subgroupsWe collected more than 4000 rental information and morethan 3000 residential house price information through thewebsite (httpsszfanglianjiacom) which cover all areas inShenzhen Table 1 lists the detailed information note thatthe price and rent in the table refer to the formal housenot the urban village In each area the rent price of theurban village is the lowest Figure 1 is a map of Shenzhenand distributions of urban village lands and nonurban villagelands in 2011

Futian Luohu and Nanshan are the core areas of Shen-zhen Investments have been made in the service industryand high-technology companies in the three areas so the

6 Journal of Advanced Transportation

Table 1 Average house price and average rental price of ten districts in Shenzhen in 2017

Spatial Location District Average house pricein 2017 (yuanm2)

Average rent in 2017(yuanm2)

Core areaFutian 52968 1195Luohu 38143 947Nanshan 56597 1217

Subcenter

Baoan 22580 769Longgang 27567 503Longhua 36432 612Yantian 33970 594

SuburbDapeng 13377 273Pingshan 12601 393

Guangming 10278 246

LegendFormal residences of NUVUrban village residences of UVGCore AreaSubcenterSuburb

0 40

Kilometers

N

Figure 1 Distributions of urban village lands and nonurban village lands in 2011

house price and rental price are the highest Baoan LonghuaLonggang and Yantian are the subcenter of Shenzhen andmanufacturing industry provides many jobs in these threeregions so the house price and rental price are lower thanin core areas The remaining three regions are relatively farfrom the city center and they have the lowest house price andrental price For core areas subcenters and suburbs there is aclear difference between house prices and rents so NUV andUVG are each divided into three subgroupsThe descriptionsare shown in Table 2

42 The Transit of Different Periods This study mainlyanalyzes the changes in public transport opportunities in thethree periods from 2011 to 2020 2011 is a base scenario and2020 is analyzed as a planning scenario Calculating trans-portation opportunity requires transportation network andactive location information Public transport data includes

subway and bus network as shown in Figure 2 The data andoperation information of subway in 2011 and 2016 come fromthe website (httpwwwszmcnet) of ShenzhenMetro GroupCo Ltd and the data and operation information of subwayin 2020 is provided by SUPB Due to the limitations of dataacquisition we use bus network and operation informationof 2016 provided by Shenzhen Urban Transport PlanningCenter (SUTPC) in three measured years We used the samelevel of bus service for the 10-year analysis period and thisoperation would cause deviations in the calculation resultsof transit opportunities This is one of the limitations ofour study The bus data contains 1700 routes and morethan 8000 bus stops Compared with the subway systemthe changes in bus services are relatively small and weassumed that using the same bus data had a smaller impacton the trends of the transit equity As seen in Figure 2subway lines are mainly concentrated in the core areas Thecoverage of subway services in suburb areas will be improved

Journal of Advanced Transportation 7

Table 2 Subgroups of NUV and UVG in Shenzhen

Group Subgroup Description Population of 2011 Percent

NUV

1 Residents who live in formal urban housing at core area(Futian Luohu and Nanshan) 4007619 278

2 Residents who live in formal urban housing atsubcenter (Longhua Yantian Longgang and Baoan) 3505463 2369

3 Residents who live in formal urban housing at suburb(Dapeng Pingshan and Guangming)

162948 037

Total 7676030 5186

UVG

4 Residents who live in the urban village at core area(Futian Luohu and Nanshan) 1402041 948

5 Residents who live in the urban village of subcenter(Longhua Yantian Longgang and Baoan)

4724083 3193

6 Residents who live in the urban village at suburb(Dapeng Pingshan and Guangming)

995467 673

Total 7121591 4814

LegendMetro line 2011Metro line 2016Metro line 2020Bus line 2016Core AreaSubcenterSuburb

0 40

Kilometers

N

Figure 2 Maps of the transit network in three periods

until 2020 so it is essential to analyze the impact of publictransport networks on fairness When calculating potentialopportunity of residents our definition of opportunity refersto jobs UPDIS provides a detailed building data of Shenzhenso the floor area of building in employment site is calculatedto represent the opportunity

5 Results and Discussions

51 The Average Transit Opportunity of Two Groups in EachDistrict The minimum unit for calculating transit oppor-tunity is the residential land unit and transit opportunityat the different aggregate levels are calculated by populationweights Figure 3 shows transit opportunity for the differentyears in Shenzhen at the community level These maps

illustrate how varied are the distributions of transit opportu-nity among the region by public bus and metro

We examined the average transit opportunity of twogroups in each district Table 3 shows the average transitopportunity of two groups at the region level Shenzhenrsquos corearea has the most significant opportunity Futian which is thecity center has the highest transit opportunities Luohu is theformer city center in the 1990s and has the second highestaccessibility Nanshan is the critical development area in thefuture with the third highest transit opportunities Thesethree regions are spatially adjacent and possess the mostpublic transport resources which have many subway linesand bus routes Baoan Longhua and Longgang are locatedoutside the core area which is the subcenter of ShenzhenThe transit opportunity in these areas is about one-third

8 Journal of Advanced Transportation

Table 3 Transit opportunity of two groups in each region

Year 2011 2016 2020NUV UVG NUV UVG NUV UVG

Futian 12381 13156 19592 21091 23095 23848Luohu 12477 13041 16143 15106 18823 17035Nanshan 5713 5886 8056 7372 10231 9035Baoan 3367 2012 3821 2185 4197 2323Longgang 4602 3337 4919 3543 6340 4820Longhua 4499 2736 4805 2848 6170 4572Yantian 1630 1285 2030 1411 4931 2858Dapeng 274 373 431 420 431 420Pingshan 1163 1107 1231 1134 1231 1134Guangming 1862 909 1881 927 2835 1297

Transit opportunity of 2011

Transit opportunity of 2020

Transit opportunity of 2016

0 40

N

Km

Transit Opportunity0000000 - 25000000002500000001 - 50000000005000000001 - 1000000000010000000001 - 1500000000015000000001 - 2000000000020000000001 - 3000000000030000000001 - 4000000000040000000001 - 80000000000

Figure 3 Maps of transit opportunity in three periods

of the core area Yantian District is a natural scenic touristarea and port area Although it is located in the subcentralarea its public transportation system is not well-developedso the transit opportunities are lower than other subcentersThe rest regions are in the outer suburbs of Shenzhenwhich are far from the core area with a small populationand minimal opportunity The opportunity of all regions isgrowing over time for both groups Transit opportunity ofcore area (especially in Futian) proliferated more than 25from 2011 to 2016 Subcenters and suburbs had a smallergrowth lower than 10 Meanwhile the differences in theabsolute value of transit opportunity between the core areasubcenter area and the suburbs were unusually large and willbe more significant in 2020 It indicates that the residents inthe city center have the most significant benefit from publictransport system

Table 4 shows the average transit opportunity in thewhole city In each measure year transit opportunity of NUVis greater than UVG The transit opportunity of UVG willincrease 45 from 2011 to 2020 and NUV will increase 63at the same time The difference in growth rate is 22When comparing the absolute value of transit opportunitythe difference between the two groups has almost doubledfrom 2011 to 2020 in 2011 NUVrsquos transit opportunity was 3279more than UVGrsquos and the gap will be 6102 in 2020 Fromthe perspective of the city NUV has more public transportadvantages than UVG For subgroups it is clear that thecore area has the highest transit opportunities whether itis UVG or NUV Subgroup 4 (residents who live in theurban village at core area) even has higher opportunities thansubgroup 2 (residents who live in formal urban housing atsubcenter)

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Journal of Advanced Transportation 3

Accessibility to opportunities is related to cumulative-opportunity and potentialgravity measures which sum thenumber of destinationsjobs reachable within certain timesby transport mode substantial literature discusses measure[26] This measure is particularly useful in describing howwell transportation networks perform about the distributionof destinations and the needs for subgroups Measure acces-sibility to opportunities usually requires an attractivenessindicator a transport network and an impedance function oftravel costThe attractiveness indicator is expressed regardingthe number of jobs or variables that represents opportunitysizeThe transport network is used to obtain travel costs andtravel cost can be travel time distance fares or a combinationof the three [27 28] There are many ways to get travelcosts such as simulation of the transport network in trafficsoftware calculation from simplified transport network inGeographic Information System (GIS) or calculation froma realistic transport network information database Theimpedance function is an inverse function which indicatesthe increase in travel costs will reduce the opportunities ofattraction

222 The Categorizing People in Transportation EquityEquity analysis needs to define a unit that can be distin-guished and units usually are groups of peoplehouseholds orregions Many studies use demographic and geographic fac-tors categorizing people to identify transport disadvantagedpeople in equity evaluation [17] and these factors includeincome car ownership age gender career household com-position and location of residence Most papers distinguishbetween high income and low income as well as car ownersand car-less individuals or households Indeed disadvan-taged status ismultidimensional some studies combine thesefactors to determine whether a person or area is a transportdisadvantaged

223 The Equity Measures in Transportation Equity Well-known horizontal equity measures are the Gini coefficientTheil index and coefficient of variation they are expressed asratios which are compared among groups to measure equityperformanceGini coefficient initially indicates level of equal-ity of income distributions in economic studies it rangesfrom0 to 1 0means absolute equality and 1 indicates absoluteinequality in transportation equity it is used to evaluate thedegree of accessibility concentration level of different regionsor groups of people and compare the level of equity beforeand after implementing a policy or transport infrastructureSome argue that the Gini coefficient fails to indicate thestructure of inequality the same Gini coefficients can havedifferent income distributions by the group Theil index isusing the information entropy concept to measure individualor interregional income inequality named The Theil indexhas good decomposability as a measure of inequality whenthe sample is divided into multiple groups The Theil indexcan measure the contribution of the intragroup gap andthe intergroup gap to the total gap so it provides moreinterpretation of the inequality among different groups Theprimary approach of vertical equity is to evaluate transport

policy or infrastructure projects according to how they affectaccessibility between disadvantaged peoplehouseholds orregions It is fairer if transportation disadvantaged groupbenefits like transport service improvements favor lower-income areas and groups or transportation services providemore access to job opportunities and other ldquobasicrdquo activi-ties

The keys to equity analysis of public transportationare the measure of accessibility and the equity measureof distribution Our work complements previous researchfrom four aspects First when evaluating the equity impactof public transport only one mode of public transportis concerned in previous studies so the result of equityevaluation could be bias We combined subways and busesto consider equity issues in this paper and proposed anenhanced potential opportunity model to measure residentsrsquobus and subway accessibility considering public servicereliability attractiveness and frequency Second due to thelimitations of data acquisition scholars discuss the impacton equity of transport policy or infrastructure projectsduring a relatively short period (before and after the imple-mentation of the target) Our research used a long-termdata to examine equity situation dynamic change and itwas helpful to capture the trends of equity influence ondifferent groups in the development of public transportThird indicators of transport distribution effects were furtherexplored and applied We used the Dagum Gini coefficientdecomposition which is more convenient thanTheil measureand kernel density estimation method to investigate thefair distribution of potential opportunities At last existingresearch focused on Europe and the United States and wetook the China fastest urbanizing city Shenzhen as a casestudy to assess transport equity by comparing accessibilityamong social groups The results can provide a referencefor the study of the impact of transportation equity in theworld

3 Methodology

In this study the public transport system was including busand subway and we divided the residents into two groupsurban village population and nonurban village populationwhich will be explained in Section 4 We used an enhancedpotential opportunity model to measure residentsrsquo bus andsubway accessibility and summarized them as transit oppor-tunity Then we used the Dagum Gini coefficient decompo-sition and kernel density estimation method to explore thefair distribution of transit opportunity among groups anddistricts from 2011 to 2020

31 Measure of Accessibility This research adopted cumu-lative-opportunity and potentialgravity measures modelsto measure transit-based job accessibility and made someenhancements Given that accessibility measurement isespecially important for the analysis of traffic equity thissection will detail how this research calculates accessibil-ity

4 Journal of Advanced Transportation

Step 1 Calculate the service range of each transit stop Theservice radius of the bus stop is 500 meters and the serviceradius of the metro station is 700 meters

Step 2 Calculate the population in each transit stop servicearea and calculate the job opportunity in each transit stopservice area In our study job opportunity is represented bythe floor area of factory company and government office

Step 3 For transit line 119897 consider both 119894 and 119895 are transit stopsof line 119897 and calculate 119860 119894119895119897 the job opportunity of 119894 can beassessed at 119895 as follows

Calculate per capita service frequency

119878119894119895119897 = 119881119894119895119897119880119875119894 (1)

119881119894119895119897 is average of vehicles (one day or a week) of transitline l from 119894 to 119895 U is transit vehicle capacity and 119875119894 is thepopulation of transit stop 119894 119878119894119895119897 reflects different transit servicecapabilities of bus and subway

Calculate 119879119894119895119897 travel time from 119894 to 119895119879119894119895119897 = 119879119886119888119888119890119904119904 + 119879119908119886119894119905 + 119879119894119899minusV119890ℎ119894119888119897119890 + 119879119890119892119903119890119904119904 (2)

The access and egress times are assumed to be 5 minof walking time which transit users are generally willing toundertake waiting time at transit stops is assumed to be one-half of the scheduled headway when the average headway oftransit service is around 10min and the in-vehicle travel timeis calculated using the scheduled arrival and departure timethat is obtained from transit service schedules

Calculate the distance decay factor

119891119894119895119897 = 11 + 120572119890minus120573119879119894119895119897 (3)

120572 and 120573 are two coefficients which need to be calibratedand we used an average travel time survey The survey datashow that in Shenzhen 963 of the people make their worktrips in less than 60 min Therefore we assumed that theconnectivity from an origin to a destination that takes 60mintravel time would be 0037 The estimates of the parametersare 120572 = 00024321 and 120573 = -0143161

Calculate the job opportunity of 119894 which can be assessedat 119895

119860 119894119895119897 = 119878119894119895119897119874119895119891119894119895119897 (4)

where Oj is the opportunity at 119895Step 4 Consider the job opportunity of 119894 can be assessed at 119890that requires a transfer between transit lines 119897 and119898

119860119896119894119890 = 12 (

119860 119894119896119897119891119894119896119897 +

119860119896119890119898119891119896119890119898 )119891

119896119894119890 (5)

119891119896119894119890 = 11 + 120572119890minus120573(119879119894119896119897+20+119879119896119890119898) (6)

where 119894 is the transit stop of line 119897 119890 is the transit stop ofline119898 and 119896 is the transfer center between 119897 and 119898

Step 5 Calculate the cumulative opportunities of i

119860 119894 = sum119895

119860 119894119895119897 +sum119896

sum119890

119860119896119894119890 (7)

Step 6 Convert residential area to the centroid and calculatethe sum of job opportunities of bus stops in the 500-meterbuffer of residential centroid and the number of metro jobopportunities of metro stations in the 700-meter buffer ofresidential centroid (if there are more than 1 transit stopsbelonging to the same line in the buffer they will be averagedas 1 transit stop) The sum of bus and metro opportunities isthe transit opportunity for residential areas

32 Decomposition of the Gini Coefficient The method ofdecomposition of the Gini coefficient in a discrete space isproposed by Dagum [29] The Dagum Gini coefficient iscalculated using the following

119866 = sum119896119895=1sum119896ℎ=2sum119899119895119894=1 sum119899ℎ119903=1 10038161003816100381610038161003816119910119895119894 minus 119910ℎ1199031003816100381610038161003816100381621198992119910 (8)

119866 is the Gini coefficient 119910119895119894(119910ℎ119903) is individual income of119894(119903) belong to subgroup 119895(ℎ) 119910 is the average income of allpopulation 119899 is the number of all population k is the numberof subgroups and 119899119895( nh) is the number of people in thej(h) subgroup Dagum decomposes the Gini coefficient intothree components [26] (1) 119866119908 contribution ofwithin groupsincome inequalities to 119866 (2) 119866119887 the net contribution of thebetween-group inequalities to119866measured on all population(3) 119866119905 the contribution of the transvariation between thesubpopulations to 119866 Detailed derivation and calculationprocess for each component are listed in [29] the equation ofGini coefficient decomposition in three components is shownas follows

119866 = 119866119908 + 119866119887 + 119866119905 (9)

Decomposition of the Gini coefficient not only effectivelysolves the source of group disparities but also describesthe distribution of subgroups and solves the problem ofoverlap between groups (shows the structure of inequality)In this paper the Gini coefficient of transit opportunity iscalculated and decomposed the population in Shenzhen isdivided into different groups it helps us to know if the transitopportunity gaps within groups generate the inequalities orif the transit opportunity gaps between groups engender theinequalities

33 Kernel Density Estimation Kernel density estimation(KDE) is a nonparametric way to estimate the probabilitydensity function of a random variable in statistics [30] Let(x1 x2 x119899) be a univariate independent and identicallydistributed sample drawn from some distribution with anunknown density Its kernel density estimator is shown asfollows

119891ℎ (119909) = 1119899ℎ119899

sum119894=1

119870(119909 minus 119909119894ℎ ) (10)

Journal of Advanced Transportation 5

where K is the kernel a nonnegative function thatintegrates to one the normal kernel is often used h gt 0 isa smoothing parameter called the bandwidth

As mentioned in Section 22 Gini coefficient or Theilmeasure are ratios which are compared among groups tomeasure equity KDE enables us to shape the distribution of avariable and analyze differences from the perspective of visualgraphics comparison So we use KDE as a complementaryway to draw probability distribution curves of transit oppor-tunity of different population groups and grasp disparity anddistribution of transit opportunity changes in decade-longurban developments in Shenzhen

4 Study Area

41 The Social Group Shenzhen is located in the Pearl RiverDelta region with a land area of 19968 km2 and an urbanpopulation of over 14 million in 2016 It is the first SpecialEconomic Zone (SEZ) city after the institution of reform andthe Open-Door Policy in China in 1979 In the past 30 yearsthe operation of a market economy has made Shenzhenrsquoseconomy develop rapidly bringing with it a dramatic popula-tion increase and spatial expansion In the study of transportequity an important part is to group residents accordingto their socioeconomic level In Shenzhen detailed data onresidentsrsquo occupations and income is not easily accessible sowe use three characteristics of a residentrsquos residence to reflecthisher socioeconomic status These three characteristics areaverage house price of residence the average rental price ofresidence and whether the residence is in the urban villagethe third feature especially is an essential basis for judging aresident as disadvantaged people

Urban village (Cheng Zhong Cun in Chinese) somescholars preferring the term ldquourbanized villagesrdquo or ldquovil-lages in the cityrdquo to avoid the confusion with the West-ern planning concept ldquourban villagerdquo is an outcome ofChinarsquos rapid urbanization and its associated rural-urbanmigration When urban expansion encroaches into ruralland the city government needs to acquire land rightsfrom the rural collective to convert the rural land intourban land The city government only expropriates thefarmland of the village to avoid the costly compensationto relocate villagers and the housing land remains in thehands of the collective Over time the village settlement issurrounded by urban built-up area creating the so-calledurban village The bond between the urban village and thedisadvantaged people is mainly caused by the residenceregistration (hukou) system and rural-urban migrationtide in China The hukou system divides the populationinto the rural population and the urban population Changefrom the rural to urban population needs to be approvedby the authorities Rural migrants can stay in large cities astemporary residents (do not have a local urban hukou) andthey are excluded from the formal urban housing marketCommercial housing is generally expensive and thus notaffordable to migrants who are employed in low-paid jobsMore affordable units provided by urban housing provisionsystem generally require a local urban hukou and are thus

not available to rural migrants Chinarsquos land policies haveenabled the native farmers in the urban villages to constructinexpensive housing units and rent out these units to the ruralmigrants In many cities the urban village is a significanttype of settlement for both local landless peasants andmigrants which are two groups with high urban povertyincidence

The data supplied by the Shenzhen Urban PlanningBureau (SUPB) and the Urban Planning and Design Instituteof Shenzhen (UPDIS) shows that there are 2942 urban villageresidential lands and 4683 nonurban village residential landsin Shenzhen 2009 The Municipal Building Survey 2009provides information for all buildings in Shenzhen includingthe urban villages There are 615702 buildings in 2009 and333576 (54) in urban villages Urban villages in Shenzhenwhich are thought to accommodate approximately sevenmillion meet the basic needs of people particularly poor andlow-income residents [31 32] There are also some studiesproving that immigrants in urban villages are vulnerableConcerning economic sectors the largest proportion ofmigrants is employed in retail hotel catering and otherservices (508)The second most common category is man-ufacturing (193) and construction (92) The proportionof people employed by highly paid public and finance sectorsis tiny [32] With the relatively poor employment profileincome among migrants is low in comparison with the cityaverage In 2004 the Municipal Government found that theaveragemonthly income amongmigrants was only 1149 yuanfar below the average personal income in the city (2195 yuan)(Shenzhen Municipal Government Housing System ReformOffice 2004)

Our study first divided the population into two groupsresidents living in urban villages called the urban villagegroup (UVG) and residents not living in urban villages calledthe Not urban village group (NUV) The UVG is mainlycomposed of low-income migrants and contains some high-income local residents A study shows that the ratio oflocal residents to migrants in the urban village is 188 [32]Thereforemost of the residents in the urban village group arelow-income people and they are more dependent on transitto access jobs and social activity The residents in NUV live informal urban houses meaning that they have owned a houseor can afford the higher rental price (the rent of a formalhouse is 25-5 times the rent in the urban village in Shenzhen)so they have higher disposable income than UVG Secondaccording to housing prices and rents in different districtsof Shenzhen each group is further divided into subgroupsWe collected more than 4000 rental information and morethan 3000 residential house price information through thewebsite (httpsszfanglianjiacom) which cover all areas inShenzhen Table 1 lists the detailed information note thatthe price and rent in the table refer to the formal housenot the urban village In each area the rent price of theurban village is the lowest Figure 1 is a map of Shenzhenand distributions of urban village lands and nonurban villagelands in 2011

Futian Luohu and Nanshan are the core areas of Shen-zhen Investments have been made in the service industryand high-technology companies in the three areas so the

6 Journal of Advanced Transportation

Table 1 Average house price and average rental price of ten districts in Shenzhen in 2017

Spatial Location District Average house pricein 2017 (yuanm2)

Average rent in 2017(yuanm2)

Core areaFutian 52968 1195Luohu 38143 947Nanshan 56597 1217

Subcenter

Baoan 22580 769Longgang 27567 503Longhua 36432 612Yantian 33970 594

SuburbDapeng 13377 273Pingshan 12601 393

Guangming 10278 246

LegendFormal residences of NUVUrban village residences of UVGCore AreaSubcenterSuburb

0 40

Kilometers

N

Figure 1 Distributions of urban village lands and nonurban village lands in 2011

house price and rental price are the highest Baoan LonghuaLonggang and Yantian are the subcenter of Shenzhen andmanufacturing industry provides many jobs in these threeregions so the house price and rental price are lower thanin core areas The remaining three regions are relatively farfrom the city center and they have the lowest house price andrental price For core areas subcenters and suburbs there is aclear difference between house prices and rents so NUV andUVG are each divided into three subgroupsThe descriptionsare shown in Table 2

42 The Transit of Different Periods This study mainlyanalyzes the changes in public transport opportunities in thethree periods from 2011 to 2020 2011 is a base scenario and2020 is analyzed as a planning scenario Calculating trans-portation opportunity requires transportation network andactive location information Public transport data includes

subway and bus network as shown in Figure 2 The data andoperation information of subway in 2011 and 2016 come fromthe website (httpwwwszmcnet) of ShenzhenMetro GroupCo Ltd and the data and operation information of subwayin 2020 is provided by SUPB Due to the limitations of dataacquisition we use bus network and operation informationof 2016 provided by Shenzhen Urban Transport PlanningCenter (SUTPC) in three measured years We used the samelevel of bus service for the 10-year analysis period and thisoperation would cause deviations in the calculation resultsof transit opportunities This is one of the limitations ofour study The bus data contains 1700 routes and morethan 8000 bus stops Compared with the subway systemthe changes in bus services are relatively small and weassumed that using the same bus data had a smaller impacton the trends of the transit equity As seen in Figure 2subway lines are mainly concentrated in the core areas Thecoverage of subway services in suburb areas will be improved

Journal of Advanced Transportation 7

Table 2 Subgroups of NUV and UVG in Shenzhen

Group Subgroup Description Population of 2011 Percent

NUV

1 Residents who live in formal urban housing at core area(Futian Luohu and Nanshan) 4007619 278

2 Residents who live in formal urban housing atsubcenter (Longhua Yantian Longgang and Baoan) 3505463 2369

3 Residents who live in formal urban housing at suburb(Dapeng Pingshan and Guangming)

162948 037

Total 7676030 5186

UVG

4 Residents who live in the urban village at core area(Futian Luohu and Nanshan) 1402041 948

5 Residents who live in the urban village of subcenter(Longhua Yantian Longgang and Baoan)

4724083 3193

6 Residents who live in the urban village at suburb(Dapeng Pingshan and Guangming)

995467 673

Total 7121591 4814

LegendMetro line 2011Metro line 2016Metro line 2020Bus line 2016Core AreaSubcenterSuburb

0 40

Kilometers

N

Figure 2 Maps of the transit network in three periods

until 2020 so it is essential to analyze the impact of publictransport networks on fairness When calculating potentialopportunity of residents our definition of opportunity refersto jobs UPDIS provides a detailed building data of Shenzhenso the floor area of building in employment site is calculatedto represent the opportunity

5 Results and Discussions

51 The Average Transit Opportunity of Two Groups in EachDistrict The minimum unit for calculating transit oppor-tunity is the residential land unit and transit opportunityat the different aggregate levels are calculated by populationweights Figure 3 shows transit opportunity for the differentyears in Shenzhen at the community level These maps

illustrate how varied are the distributions of transit opportu-nity among the region by public bus and metro

We examined the average transit opportunity of twogroups in each district Table 3 shows the average transitopportunity of two groups at the region level Shenzhenrsquos corearea has the most significant opportunity Futian which is thecity center has the highest transit opportunities Luohu is theformer city center in the 1990s and has the second highestaccessibility Nanshan is the critical development area in thefuture with the third highest transit opportunities Thesethree regions are spatially adjacent and possess the mostpublic transport resources which have many subway linesand bus routes Baoan Longhua and Longgang are locatedoutside the core area which is the subcenter of ShenzhenThe transit opportunity in these areas is about one-third

8 Journal of Advanced Transportation

Table 3 Transit opportunity of two groups in each region

Year 2011 2016 2020NUV UVG NUV UVG NUV UVG

Futian 12381 13156 19592 21091 23095 23848Luohu 12477 13041 16143 15106 18823 17035Nanshan 5713 5886 8056 7372 10231 9035Baoan 3367 2012 3821 2185 4197 2323Longgang 4602 3337 4919 3543 6340 4820Longhua 4499 2736 4805 2848 6170 4572Yantian 1630 1285 2030 1411 4931 2858Dapeng 274 373 431 420 431 420Pingshan 1163 1107 1231 1134 1231 1134Guangming 1862 909 1881 927 2835 1297

Transit opportunity of 2011

Transit opportunity of 2020

Transit opportunity of 2016

0 40

N

Km

Transit Opportunity0000000 - 25000000002500000001 - 50000000005000000001 - 1000000000010000000001 - 1500000000015000000001 - 2000000000020000000001 - 3000000000030000000001 - 4000000000040000000001 - 80000000000

Figure 3 Maps of transit opportunity in three periods

of the core area Yantian District is a natural scenic touristarea and port area Although it is located in the subcentralarea its public transportation system is not well-developedso the transit opportunities are lower than other subcentersThe rest regions are in the outer suburbs of Shenzhenwhich are far from the core area with a small populationand minimal opportunity The opportunity of all regions isgrowing over time for both groups Transit opportunity ofcore area (especially in Futian) proliferated more than 25from 2011 to 2016 Subcenters and suburbs had a smallergrowth lower than 10 Meanwhile the differences in theabsolute value of transit opportunity between the core areasubcenter area and the suburbs were unusually large and willbe more significant in 2020 It indicates that the residents inthe city center have the most significant benefit from publictransport system

Table 4 shows the average transit opportunity in thewhole city In each measure year transit opportunity of NUVis greater than UVG The transit opportunity of UVG willincrease 45 from 2011 to 2020 and NUV will increase 63at the same time The difference in growth rate is 22When comparing the absolute value of transit opportunitythe difference between the two groups has almost doubledfrom 2011 to 2020 in 2011 NUVrsquos transit opportunity was 3279more than UVGrsquos and the gap will be 6102 in 2020 Fromthe perspective of the city NUV has more public transportadvantages than UVG For subgroups it is clear that thecore area has the highest transit opportunities whether itis UVG or NUV Subgroup 4 (residents who live in theurban village at core area) even has higher opportunities thansubgroup 2 (residents who live in formal urban housing atsubcenter)

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

4 Journal of Advanced Transportation

Step 1 Calculate the service range of each transit stop Theservice radius of the bus stop is 500 meters and the serviceradius of the metro station is 700 meters

Step 2 Calculate the population in each transit stop servicearea and calculate the job opportunity in each transit stopservice area In our study job opportunity is represented bythe floor area of factory company and government office

Step 3 For transit line 119897 consider both 119894 and 119895 are transit stopsof line 119897 and calculate 119860 119894119895119897 the job opportunity of 119894 can beassessed at 119895 as follows

Calculate per capita service frequency

119878119894119895119897 = 119881119894119895119897119880119875119894 (1)

119881119894119895119897 is average of vehicles (one day or a week) of transitline l from 119894 to 119895 U is transit vehicle capacity and 119875119894 is thepopulation of transit stop 119894 119878119894119895119897 reflects different transit servicecapabilities of bus and subway

Calculate 119879119894119895119897 travel time from 119894 to 119895119879119894119895119897 = 119879119886119888119888119890119904119904 + 119879119908119886119894119905 + 119879119894119899minusV119890ℎ119894119888119897119890 + 119879119890119892119903119890119904119904 (2)

The access and egress times are assumed to be 5 minof walking time which transit users are generally willing toundertake waiting time at transit stops is assumed to be one-half of the scheduled headway when the average headway oftransit service is around 10min and the in-vehicle travel timeis calculated using the scheduled arrival and departure timethat is obtained from transit service schedules

Calculate the distance decay factor

119891119894119895119897 = 11 + 120572119890minus120573119879119894119895119897 (3)

120572 and 120573 are two coefficients which need to be calibratedand we used an average travel time survey The survey datashow that in Shenzhen 963 of the people make their worktrips in less than 60 min Therefore we assumed that theconnectivity from an origin to a destination that takes 60mintravel time would be 0037 The estimates of the parametersare 120572 = 00024321 and 120573 = -0143161

Calculate the job opportunity of 119894 which can be assessedat 119895

119860 119894119895119897 = 119878119894119895119897119874119895119891119894119895119897 (4)

where Oj is the opportunity at 119895Step 4 Consider the job opportunity of 119894 can be assessed at 119890that requires a transfer between transit lines 119897 and119898

119860119896119894119890 = 12 (

119860 119894119896119897119891119894119896119897 +

119860119896119890119898119891119896119890119898 )119891

119896119894119890 (5)

119891119896119894119890 = 11 + 120572119890minus120573(119879119894119896119897+20+119879119896119890119898) (6)

where 119894 is the transit stop of line 119897 119890 is the transit stop ofline119898 and 119896 is the transfer center between 119897 and 119898

Step 5 Calculate the cumulative opportunities of i

119860 119894 = sum119895

119860 119894119895119897 +sum119896

sum119890

119860119896119894119890 (7)

Step 6 Convert residential area to the centroid and calculatethe sum of job opportunities of bus stops in the 500-meterbuffer of residential centroid and the number of metro jobopportunities of metro stations in the 700-meter buffer ofresidential centroid (if there are more than 1 transit stopsbelonging to the same line in the buffer they will be averagedas 1 transit stop) The sum of bus and metro opportunities isthe transit opportunity for residential areas

32 Decomposition of the Gini Coefficient The method ofdecomposition of the Gini coefficient in a discrete space isproposed by Dagum [29] The Dagum Gini coefficient iscalculated using the following

119866 = sum119896119895=1sum119896ℎ=2sum119899119895119894=1 sum119899ℎ119903=1 10038161003816100381610038161003816119910119895119894 minus 119910ℎ1199031003816100381610038161003816100381621198992119910 (8)

119866 is the Gini coefficient 119910119895119894(119910ℎ119903) is individual income of119894(119903) belong to subgroup 119895(ℎ) 119910 is the average income of allpopulation 119899 is the number of all population k is the numberof subgroups and 119899119895( nh) is the number of people in thej(h) subgroup Dagum decomposes the Gini coefficient intothree components [26] (1) 119866119908 contribution ofwithin groupsincome inequalities to 119866 (2) 119866119887 the net contribution of thebetween-group inequalities to119866measured on all population(3) 119866119905 the contribution of the transvariation between thesubpopulations to 119866 Detailed derivation and calculationprocess for each component are listed in [29] the equation ofGini coefficient decomposition in three components is shownas follows

119866 = 119866119908 + 119866119887 + 119866119905 (9)

Decomposition of the Gini coefficient not only effectivelysolves the source of group disparities but also describesthe distribution of subgroups and solves the problem ofoverlap between groups (shows the structure of inequality)In this paper the Gini coefficient of transit opportunity iscalculated and decomposed the population in Shenzhen isdivided into different groups it helps us to know if the transitopportunity gaps within groups generate the inequalities orif the transit opportunity gaps between groups engender theinequalities

33 Kernel Density Estimation Kernel density estimation(KDE) is a nonparametric way to estimate the probabilitydensity function of a random variable in statistics [30] Let(x1 x2 x119899) be a univariate independent and identicallydistributed sample drawn from some distribution with anunknown density Its kernel density estimator is shown asfollows

119891ℎ (119909) = 1119899ℎ119899

sum119894=1

119870(119909 minus 119909119894ℎ ) (10)

Journal of Advanced Transportation 5

where K is the kernel a nonnegative function thatintegrates to one the normal kernel is often used h gt 0 isa smoothing parameter called the bandwidth

As mentioned in Section 22 Gini coefficient or Theilmeasure are ratios which are compared among groups tomeasure equity KDE enables us to shape the distribution of avariable and analyze differences from the perspective of visualgraphics comparison So we use KDE as a complementaryway to draw probability distribution curves of transit oppor-tunity of different population groups and grasp disparity anddistribution of transit opportunity changes in decade-longurban developments in Shenzhen

4 Study Area

41 The Social Group Shenzhen is located in the Pearl RiverDelta region with a land area of 19968 km2 and an urbanpopulation of over 14 million in 2016 It is the first SpecialEconomic Zone (SEZ) city after the institution of reform andthe Open-Door Policy in China in 1979 In the past 30 yearsthe operation of a market economy has made Shenzhenrsquoseconomy develop rapidly bringing with it a dramatic popula-tion increase and spatial expansion In the study of transportequity an important part is to group residents accordingto their socioeconomic level In Shenzhen detailed data onresidentsrsquo occupations and income is not easily accessible sowe use three characteristics of a residentrsquos residence to reflecthisher socioeconomic status These three characteristics areaverage house price of residence the average rental price ofresidence and whether the residence is in the urban villagethe third feature especially is an essential basis for judging aresident as disadvantaged people

Urban village (Cheng Zhong Cun in Chinese) somescholars preferring the term ldquourbanized villagesrdquo or ldquovil-lages in the cityrdquo to avoid the confusion with the West-ern planning concept ldquourban villagerdquo is an outcome ofChinarsquos rapid urbanization and its associated rural-urbanmigration When urban expansion encroaches into ruralland the city government needs to acquire land rightsfrom the rural collective to convert the rural land intourban land The city government only expropriates thefarmland of the village to avoid the costly compensationto relocate villagers and the housing land remains in thehands of the collective Over time the village settlement issurrounded by urban built-up area creating the so-calledurban village The bond between the urban village and thedisadvantaged people is mainly caused by the residenceregistration (hukou) system and rural-urban migrationtide in China The hukou system divides the populationinto the rural population and the urban population Changefrom the rural to urban population needs to be approvedby the authorities Rural migrants can stay in large cities astemporary residents (do not have a local urban hukou) andthey are excluded from the formal urban housing marketCommercial housing is generally expensive and thus notaffordable to migrants who are employed in low-paid jobsMore affordable units provided by urban housing provisionsystem generally require a local urban hukou and are thus

not available to rural migrants Chinarsquos land policies haveenabled the native farmers in the urban villages to constructinexpensive housing units and rent out these units to the ruralmigrants In many cities the urban village is a significanttype of settlement for both local landless peasants andmigrants which are two groups with high urban povertyincidence

The data supplied by the Shenzhen Urban PlanningBureau (SUPB) and the Urban Planning and Design Instituteof Shenzhen (UPDIS) shows that there are 2942 urban villageresidential lands and 4683 nonurban village residential landsin Shenzhen 2009 The Municipal Building Survey 2009provides information for all buildings in Shenzhen includingthe urban villages There are 615702 buildings in 2009 and333576 (54) in urban villages Urban villages in Shenzhenwhich are thought to accommodate approximately sevenmillion meet the basic needs of people particularly poor andlow-income residents [31 32] There are also some studiesproving that immigrants in urban villages are vulnerableConcerning economic sectors the largest proportion ofmigrants is employed in retail hotel catering and otherservices (508)The second most common category is man-ufacturing (193) and construction (92) The proportionof people employed by highly paid public and finance sectorsis tiny [32] With the relatively poor employment profileincome among migrants is low in comparison with the cityaverage In 2004 the Municipal Government found that theaveragemonthly income amongmigrants was only 1149 yuanfar below the average personal income in the city (2195 yuan)(Shenzhen Municipal Government Housing System ReformOffice 2004)

Our study first divided the population into two groupsresidents living in urban villages called the urban villagegroup (UVG) and residents not living in urban villages calledthe Not urban village group (NUV) The UVG is mainlycomposed of low-income migrants and contains some high-income local residents A study shows that the ratio oflocal residents to migrants in the urban village is 188 [32]Thereforemost of the residents in the urban village group arelow-income people and they are more dependent on transitto access jobs and social activity The residents in NUV live informal urban houses meaning that they have owned a houseor can afford the higher rental price (the rent of a formalhouse is 25-5 times the rent in the urban village in Shenzhen)so they have higher disposable income than UVG Secondaccording to housing prices and rents in different districtsof Shenzhen each group is further divided into subgroupsWe collected more than 4000 rental information and morethan 3000 residential house price information through thewebsite (httpsszfanglianjiacom) which cover all areas inShenzhen Table 1 lists the detailed information note thatthe price and rent in the table refer to the formal housenot the urban village In each area the rent price of theurban village is the lowest Figure 1 is a map of Shenzhenand distributions of urban village lands and nonurban villagelands in 2011

Futian Luohu and Nanshan are the core areas of Shen-zhen Investments have been made in the service industryand high-technology companies in the three areas so the

6 Journal of Advanced Transportation

Table 1 Average house price and average rental price of ten districts in Shenzhen in 2017

Spatial Location District Average house pricein 2017 (yuanm2)

Average rent in 2017(yuanm2)

Core areaFutian 52968 1195Luohu 38143 947Nanshan 56597 1217

Subcenter

Baoan 22580 769Longgang 27567 503Longhua 36432 612Yantian 33970 594

SuburbDapeng 13377 273Pingshan 12601 393

Guangming 10278 246

LegendFormal residences of NUVUrban village residences of UVGCore AreaSubcenterSuburb

0 40

Kilometers

N

Figure 1 Distributions of urban village lands and nonurban village lands in 2011

house price and rental price are the highest Baoan LonghuaLonggang and Yantian are the subcenter of Shenzhen andmanufacturing industry provides many jobs in these threeregions so the house price and rental price are lower thanin core areas The remaining three regions are relatively farfrom the city center and they have the lowest house price andrental price For core areas subcenters and suburbs there is aclear difference between house prices and rents so NUV andUVG are each divided into three subgroupsThe descriptionsare shown in Table 2

42 The Transit of Different Periods This study mainlyanalyzes the changes in public transport opportunities in thethree periods from 2011 to 2020 2011 is a base scenario and2020 is analyzed as a planning scenario Calculating trans-portation opportunity requires transportation network andactive location information Public transport data includes

subway and bus network as shown in Figure 2 The data andoperation information of subway in 2011 and 2016 come fromthe website (httpwwwszmcnet) of ShenzhenMetro GroupCo Ltd and the data and operation information of subwayin 2020 is provided by SUPB Due to the limitations of dataacquisition we use bus network and operation informationof 2016 provided by Shenzhen Urban Transport PlanningCenter (SUTPC) in three measured years We used the samelevel of bus service for the 10-year analysis period and thisoperation would cause deviations in the calculation resultsof transit opportunities This is one of the limitations ofour study The bus data contains 1700 routes and morethan 8000 bus stops Compared with the subway systemthe changes in bus services are relatively small and weassumed that using the same bus data had a smaller impacton the trends of the transit equity As seen in Figure 2subway lines are mainly concentrated in the core areas Thecoverage of subway services in suburb areas will be improved

Journal of Advanced Transportation 7

Table 2 Subgroups of NUV and UVG in Shenzhen

Group Subgroup Description Population of 2011 Percent

NUV

1 Residents who live in formal urban housing at core area(Futian Luohu and Nanshan) 4007619 278

2 Residents who live in formal urban housing atsubcenter (Longhua Yantian Longgang and Baoan) 3505463 2369

3 Residents who live in formal urban housing at suburb(Dapeng Pingshan and Guangming)

162948 037

Total 7676030 5186

UVG

4 Residents who live in the urban village at core area(Futian Luohu and Nanshan) 1402041 948

5 Residents who live in the urban village of subcenter(Longhua Yantian Longgang and Baoan)

4724083 3193

6 Residents who live in the urban village at suburb(Dapeng Pingshan and Guangming)

995467 673

Total 7121591 4814

LegendMetro line 2011Metro line 2016Metro line 2020Bus line 2016Core AreaSubcenterSuburb

0 40

Kilometers

N

Figure 2 Maps of the transit network in three periods

until 2020 so it is essential to analyze the impact of publictransport networks on fairness When calculating potentialopportunity of residents our definition of opportunity refersto jobs UPDIS provides a detailed building data of Shenzhenso the floor area of building in employment site is calculatedto represent the opportunity

5 Results and Discussions

51 The Average Transit Opportunity of Two Groups in EachDistrict The minimum unit for calculating transit oppor-tunity is the residential land unit and transit opportunityat the different aggregate levels are calculated by populationweights Figure 3 shows transit opportunity for the differentyears in Shenzhen at the community level These maps

illustrate how varied are the distributions of transit opportu-nity among the region by public bus and metro

We examined the average transit opportunity of twogroups in each district Table 3 shows the average transitopportunity of two groups at the region level Shenzhenrsquos corearea has the most significant opportunity Futian which is thecity center has the highest transit opportunities Luohu is theformer city center in the 1990s and has the second highestaccessibility Nanshan is the critical development area in thefuture with the third highest transit opportunities Thesethree regions are spatially adjacent and possess the mostpublic transport resources which have many subway linesand bus routes Baoan Longhua and Longgang are locatedoutside the core area which is the subcenter of ShenzhenThe transit opportunity in these areas is about one-third

8 Journal of Advanced Transportation

Table 3 Transit opportunity of two groups in each region

Year 2011 2016 2020NUV UVG NUV UVG NUV UVG

Futian 12381 13156 19592 21091 23095 23848Luohu 12477 13041 16143 15106 18823 17035Nanshan 5713 5886 8056 7372 10231 9035Baoan 3367 2012 3821 2185 4197 2323Longgang 4602 3337 4919 3543 6340 4820Longhua 4499 2736 4805 2848 6170 4572Yantian 1630 1285 2030 1411 4931 2858Dapeng 274 373 431 420 431 420Pingshan 1163 1107 1231 1134 1231 1134Guangming 1862 909 1881 927 2835 1297

Transit opportunity of 2011

Transit opportunity of 2020

Transit opportunity of 2016

0 40

N

Km

Transit Opportunity0000000 - 25000000002500000001 - 50000000005000000001 - 1000000000010000000001 - 1500000000015000000001 - 2000000000020000000001 - 3000000000030000000001 - 4000000000040000000001 - 80000000000

Figure 3 Maps of transit opportunity in three periods

of the core area Yantian District is a natural scenic touristarea and port area Although it is located in the subcentralarea its public transportation system is not well-developedso the transit opportunities are lower than other subcentersThe rest regions are in the outer suburbs of Shenzhenwhich are far from the core area with a small populationand minimal opportunity The opportunity of all regions isgrowing over time for both groups Transit opportunity ofcore area (especially in Futian) proliferated more than 25from 2011 to 2016 Subcenters and suburbs had a smallergrowth lower than 10 Meanwhile the differences in theabsolute value of transit opportunity between the core areasubcenter area and the suburbs were unusually large and willbe more significant in 2020 It indicates that the residents inthe city center have the most significant benefit from publictransport system

Table 4 shows the average transit opportunity in thewhole city In each measure year transit opportunity of NUVis greater than UVG The transit opportunity of UVG willincrease 45 from 2011 to 2020 and NUV will increase 63at the same time The difference in growth rate is 22When comparing the absolute value of transit opportunitythe difference between the two groups has almost doubledfrom 2011 to 2020 in 2011 NUVrsquos transit opportunity was 3279more than UVGrsquos and the gap will be 6102 in 2020 Fromthe perspective of the city NUV has more public transportadvantages than UVG For subgroups it is clear that thecore area has the highest transit opportunities whether itis UVG or NUV Subgroup 4 (residents who live in theurban village at core area) even has higher opportunities thansubgroup 2 (residents who live in formal urban housing atsubcenter)

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Journal of Advanced Transportation 5

where K is the kernel a nonnegative function thatintegrates to one the normal kernel is often used h gt 0 isa smoothing parameter called the bandwidth

As mentioned in Section 22 Gini coefficient or Theilmeasure are ratios which are compared among groups tomeasure equity KDE enables us to shape the distribution of avariable and analyze differences from the perspective of visualgraphics comparison So we use KDE as a complementaryway to draw probability distribution curves of transit oppor-tunity of different population groups and grasp disparity anddistribution of transit opportunity changes in decade-longurban developments in Shenzhen

4 Study Area

41 The Social Group Shenzhen is located in the Pearl RiverDelta region with a land area of 19968 km2 and an urbanpopulation of over 14 million in 2016 It is the first SpecialEconomic Zone (SEZ) city after the institution of reform andthe Open-Door Policy in China in 1979 In the past 30 yearsthe operation of a market economy has made Shenzhenrsquoseconomy develop rapidly bringing with it a dramatic popula-tion increase and spatial expansion In the study of transportequity an important part is to group residents accordingto their socioeconomic level In Shenzhen detailed data onresidentsrsquo occupations and income is not easily accessible sowe use three characteristics of a residentrsquos residence to reflecthisher socioeconomic status These three characteristics areaverage house price of residence the average rental price ofresidence and whether the residence is in the urban villagethe third feature especially is an essential basis for judging aresident as disadvantaged people

Urban village (Cheng Zhong Cun in Chinese) somescholars preferring the term ldquourbanized villagesrdquo or ldquovil-lages in the cityrdquo to avoid the confusion with the West-ern planning concept ldquourban villagerdquo is an outcome ofChinarsquos rapid urbanization and its associated rural-urbanmigration When urban expansion encroaches into ruralland the city government needs to acquire land rightsfrom the rural collective to convert the rural land intourban land The city government only expropriates thefarmland of the village to avoid the costly compensationto relocate villagers and the housing land remains in thehands of the collective Over time the village settlement issurrounded by urban built-up area creating the so-calledurban village The bond between the urban village and thedisadvantaged people is mainly caused by the residenceregistration (hukou) system and rural-urban migrationtide in China The hukou system divides the populationinto the rural population and the urban population Changefrom the rural to urban population needs to be approvedby the authorities Rural migrants can stay in large cities astemporary residents (do not have a local urban hukou) andthey are excluded from the formal urban housing marketCommercial housing is generally expensive and thus notaffordable to migrants who are employed in low-paid jobsMore affordable units provided by urban housing provisionsystem generally require a local urban hukou and are thus

not available to rural migrants Chinarsquos land policies haveenabled the native farmers in the urban villages to constructinexpensive housing units and rent out these units to the ruralmigrants In many cities the urban village is a significanttype of settlement for both local landless peasants andmigrants which are two groups with high urban povertyincidence

The data supplied by the Shenzhen Urban PlanningBureau (SUPB) and the Urban Planning and Design Instituteof Shenzhen (UPDIS) shows that there are 2942 urban villageresidential lands and 4683 nonurban village residential landsin Shenzhen 2009 The Municipal Building Survey 2009provides information for all buildings in Shenzhen includingthe urban villages There are 615702 buildings in 2009 and333576 (54) in urban villages Urban villages in Shenzhenwhich are thought to accommodate approximately sevenmillion meet the basic needs of people particularly poor andlow-income residents [31 32] There are also some studiesproving that immigrants in urban villages are vulnerableConcerning economic sectors the largest proportion ofmigrants is employed in retail hotel catering and otherservices (508)The second most common category is man-ufacturing (193) and construction (92) The proportionof people employed by highly paid public and finance sectorsis tiny [32] With the relatively poor employment profileincome among migrants is low in comparison with the cityaverage In 2004 the Municipal Government found that theaveragemonthly income amongmigrants was only 1149 yuanfar below the average personal income in the city (2195 yuan)(Shenzhen Municipal Government Housing System ReformOffice 2004)

Our study first divided the population into two groupsresidents living in urban villages called the urban villagegroup (UVG) and residents not living in urban villages calledthe Not urban village group (NUV) The UVG is mainlycomposed of low-income migrants and contains some high-income local residents A study shows that the ratio oflocal residents to migrants in the urban village is 188 [32]Thereforemost of the residents in the urban village group arelow-income people and they are more dependent on transitto access jobs and social activity The residents in NUV live informal urban houses meaning that they have owned a houseor can afford the higher rental price (the rent of a formalhouse is 25-5 times the rent in the urban village in Shenzhen)so they have higher disposable income than UVG Secondaccording to housing prices and rents in different districtsof Shenzhen each group is further divided into subgroupsWe collected more than 4000 rental information and morethan 3000 residential house price information through thewebsite (httpsszfanglianjiacom) which cover all areas inShenzhen Table 1 lists the detailed information note thatthe price and rent in the table refer to the formal housenot the urban village In each area the rent price of theurban village is the lowest Figure 1 is a map of Shenzhenand distributions of urban village lands and nonurban villagelands in 2011

Futian Luohu and Nanshan are the core areas of Shen-zhen Investments have been made in the service industryand high-technology companies in the three areas so the

6 Journal of Advanced Transportation

Table 1 Average house price and average rental price of ten districts in Shenzhen in 2017

Spatial Location District Average house pricein 2017 (yuanm2)

Average rent in 2017(yuanm2)

Core areaFutian 52968 1195Luohu 38143 947Nanshan 56597 1217

Subcenter

Baoan 22580 769Longgang 27567 503Longhua 36432 612Yantian 33970 594

SuburbDapeng 13377 273Pingshan 12601 393

Guangming 10278 246

LegendFormal residences of NUVUrban village residences of UVGCore AreaSubcenterSuburb

0 40

Kilometers

N

Figure 1 Distributions of urban village lands and nonurban village lands in 2011

house price and rental price are the highest Baoan LonghuaLonggang and Yantian are the subcenter of Shenzhen andmanufacturing industry provides many jobs in these threeregions so the house price and rental price are lower thanin core areas The remaining three regions are relatively farfrom the city center and they have the lowest house price andrental price For core areas subcenters and suburbs there is aclear difference between house prices and rents so NUV andUVG are each divided into three subgroupsThe descriptionsare shown in Table 2

42 The Transit of Different Periods This study mainlyanalyzes the changes in public transport opportunities in thethree periods from 2011 to 2020 2011 is a base scenario and2020 is analyzed as a planning scenario Calculating trans-portation opportunity requires transportation network andactive location information Public transport data includes

subway and bus network as shown in Figure 2 The data andoperation information of subway in 2011 and 2016 come fromthe website (httpwwwszmcnet) of ShenzhenMetro GroupCo Ltd and the data and operation information of subwayin 2020 is provided by SUPB Due to the limitations of dataacquisition we use bus network and operation informationof 2016 provided by Shenzhen Urban Transport PlanningCenter (SUTPC) in three measured years We used the samelevel of bus service for the 10-year analysis period and thisoperation would cause deviations in the calculation resultsof transit opportunities This is one of the limitations ofour study The bus data contains 1700 routes and morethan 8000 bus stops Compared with the subway systemthe changes in bus services are relatively small and weassumed that using the same bus data had a smaller impacton the trends of the transit equity As seen in Figure 2subway lines are mainly concentrated in the core areas Thecoverage of subway services in suburb areas will be improved

Journal of Advanced Transportation 7

Table 2 Subgroups of NUV and UVG in Shenzhen

Group Subgroup Description Population of 2011 Percent

NUV

1 Residents who live in formal urban housing at core area(Futian Luohu and Nanshan) 4007619 278

2 Residents who live in formal urban housing atsubcenter (Longhua Yantian Longgang and Baoan) 3505463 2369

3 Residents who live in formal urban housing at suburb(Dapeng Pingshan and Guangming)

162948 037

Total 7676030 5186

UVG

4 Residents who live in the urban village at core area(Futian Luohu and Nanshan) 1402041 948

5 Residents who live in the urban village of subcenter(Longhua Yantian Longgang and Baoan)

4724083 3193

6 Residents who live in the urban village at suburb(Dapeng Pingshan and Guangming)

995467 673

Total 7121591 4814

LegendMetro line 2011Metro line 2016Metro line 2020Bus line 2016Core AreaSubcenterSuburb

0 40

Kilometers

N

Figure 2 Maps of the transit network in three periods

until 2020 so it is essential to analyze the impact of publictransport networks on fairness When calculating potentialopportunity of residents our definition of opportunity refersto jobs UPDIS provides a detailed building data of Shenzhenso the floor area of building in employment site is calculatedto represent the opportunity

5 Results and Discussions

51 The Average Transit Opportunity of Two Groups in EachDistrict The minimum unit for calculating transit oppor-tunity is the residential land unit and transit opportunityat the different aggregate levels are calculated by populationweights Figure 3 shows transit opportunity for the differentyears in Shenzhen at the community level These maps

illustrate how varied are the distributions of transit opportu-nity among the region by public bus and metro

We examined the average transit opportunity of twogroups in each district Table 3 shows the average transitopportunity of two groups at the region level Shenzhenrsquos corearea has the most significant opportunity Futian which is thecity center has the highest transit opportunities Luohu is theformer city center in the 1990s and has the second highestaccessibility Nanshan is the critical development area in thefuture with the third highest transit opportunities Thesethree regions are spatially adjacent and possess the mostpublic transport resources which have many subway linesand bus routes Baoan Longhua and Longgang are locatedoutside the core area which is the subcenter of ShenzhenThe transit opportunity in these areas is about one-third

8 Journal of Advanced Transportation

Table 3 Transit opportunity of two groups in each region

Year 2011 2016 2020NUV UVG NUV UVG NUV UVG

Futian 12381 13156 19592 21091 23095 23848Luohu 12477 13041 16143 15106 18823 17035Nanshan 5713 5886 8056 7372 10231 9035Baoan 3367 2012 3821 2185 4197 2323Longgang 4602 3337 4919 3543 6340 4820Longhua 4499 2736 4805 2848 6170 4572Yantian 1630 1285 2030 1411 4931 2858Dapeng 274 373 431 420 431 420Pingshan 1163 1107 1231 1134 1231 1134Guangming 1862 909 1881 927 2835 1297

Transit opportunity of 2011

Transit opportunity of 2020

Transit opportunity of 2016

0 40

N

Km

Transit Opportunity0000000 - 25000000002500000001 - 50000000005000000001 - 1000000000010000000001 - 1500000000015000000001 - 2000000000020000000001 - 3000000000030000000001 - 4000000000040000000001 - 80000000000

Figure 3 Maps of transit opportunity in three periods

of the core area Yantian District is a natural scenic touristarea and port area Although it is located in the subcentralarea its public transportation system is not well-developedso the transit opportunities are lower than other subcentersThe rest regions are in the outer suburbs of Shenzhenwhich are far from the core area with a small populationand minimal opportunity The opportunity of all regions isgrowing over time for both groups Transit opportunity ofcore area (especially in Futian) proliferated more than 25from 2011 to 2016 Subcenters and suburbs had a smallergrowth lower than 10 Meanwhile the differences in theabsolute value of transit opportunity between the core areasubcenter area and the suburbs were unusually large and willbe more significant in 2020 It indicates that the residents inthe city center have the most significant benefit from publictransport system

Table 4 shows the average transit opportunity in thewhole city In each measure year transit opportunity of NUVis greater than UVG The transit opportunity of UVG willincrease 45 from 2011 to 2020 and NUV will increase 63at the same time The difference in growth rate is 22When comparing the absolute value of transit opportunitythe difference between the two groups has almost doubledfrom 2011 to 2020 in 2011 NUVrsquos transit opportunity was 3279more than UVGrsquos and the gap will be 6102 in 2020 Fromthe perspective of the city NUV has more public transportadvantages than UVG For subgroups it is clear that thecore area has the highest transit opportunities whether itis UVG or NUV Subgroup 4 (residents who live in theurban village at core area) even has higher opportunities thansubgroup 2 (residents who live in formal urban housing atsubcenter)

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

6 Journal of Advanced Transportation

Table 1 Average house price and average rental price of ten districts in Shenzhen in 2017

Spatial Location District Average house pricein 2017 (yuanm2)

Average rent in 2017(yuanm2)

Core areaFutian 52968 1195Luohu 38143 947Nanshan 56597 1217

Subcenter

Baoan 22580 769Longgang 27567 503Longhua 36432 612Yantian 33970 594

SuburbDapeng 13377 273Pingshan 12601 393

Guangming 10278 246

LegendFormal residences of NUVUrban village residences of UVGCore AreaSubcenterSuburb

0 40

Kilometers

N

Figure 1 Distributions of urban village lands and nonurban village lands in 2011

house price and rental price are the highest Baoan LonghuaLonggang and Yantian are the subcenter of Shenzhen andmanufacturing industry provides many jobs in these threeregions so the house price and rental price are lower thanin core areas The remaining three regions are relatively farfrom the city center and they have the lowest house price andrental price For core areas subcenters and suburbs there is aclear difference between house prices and rents so NUV andUVG are each divided into three subgroupsThe descriptionsare shown in Table 2

42 The Transit of Different Periods This study mainlyanalyzes the changes in public transport opportunities in thethree periods from 2011 to 2020 2011 is a base scenario and2020 is analyzed as a planning scenario Calculating trans-portation opportunity requires transportation network andactive location information Public transport data includes

subway and bus network as shown in Figure 2 The data andoperation information of subway in 2011 and 2016 come fromthe website (httpwwwszmcnet) of ShenzhenMetro GroupCo Ltd and the data and operation information of subwayin 2020 is provided by SUPB Due to the limitations of dataacquisition we use bus network and operation informationof 2016 provided by Shenzhen Urban Transport PlanningCenter (SUTPC) in three measured years We used the samelevel of bus service for the 10-year analysis period and thisoperation would cause deviations in the calculation resultsof transit opportunities This is one of the limitations ofour study The bus data contains 1700 routes and morethan 8000 bus stops Compared with the subway systemthe changes in bus services are relatively small and weassumed that using the same bus data had a smaller impacton the trends of the transit equity As seen in Figure 2subway lines are mainly concentrated in the core areas Thecoverage of subway services in suburb areas will be improved

Journal of Advanced Transportation 7

Table 2 Subgroups of NUV and UVG in Shenzhen

Group Subgroup Description Population of 2011 Percent

NUV

1 Residents who live in formal urban housing at core area(Futian Luohu and Nanshan) 4007619 278

2 Residents who live in formal urban housing atsubcenter (Longhua Yantian Longgang and Baoan) 3505463 2369

3 Residents who live in formal urban housing at suburb(Dapeng Pingshan and Guangming)

162948 037

Total 7676030 5186

UVG

4 Residents who live in the urban village at core area(Futian Luohu and Nanshan) 1402041 948

5 Residents who live in the urban village of subcenter(Longhua Yantian Longgang and Baoan)

4724083 3193

6 Residents who live in the urban village at suburb(Dapeng Pingshan and Guangming)

995467 673

Total 7121591 4814

LegendMetro line 2011Metro line 2016Metro line 2020Bus line 2016Core AreaSubcenterSuburb

0 40

Kilometers

N

Figure 2 Maps of the transit network in three periods

until 2020 so it is essential to analyze the impact of publictransport networks on fairness When calculating potentialopportunity of residents our definition of opportunity refersto jobs UPDIS provides a detailed building data of Shenzhenso the floor area of building in employment site is calculatedto represent the opportunity

5 Results and Discussions

51 The Average Transit Opportunity of Two Groups in EachDistrict The minimum unit for calculating transit oppor-tunity is the residential land unit and transit opportunityat the different aggregate levels are calculated by populationweights Figure 3 shows transit opportunity for the differentyears in Shenzhen at the community level These maps

illustrate how varied are the distributions of transit opportu-nity among the region by public bus and metro

We examined the average transit opportunity of twogroups in each district Table 3 shows the average transitopportunity of two groups at the region level Shenzhenrsquos corearea has the most significant opportunity Futian which is thecity center has the highest transit opportunities Luohu is theformer city center in the 1990s and has the second highestaccessibility Nanshan is the critical development area in thefuture with the third highest transit opportunities Thesethree regions are spatially adjacent and possess the mostpublic transport resources which have many subway linesand bus routes Baoan Longhua and Longgang are locatedoutside the core area which is the subcenter of ShenzhenThe transit opportunity in these areas is about one-third

8 Journal of Advanced Transportation

Table 3 Transit opportunity of two groups in each region

Year 2011 2016 2020NUV UVG NUV UVG NUV UVG

Futian 12381 13156 19592 21091 23095 23848Luohu 12477 13041 16143 15106 18823 17035Nanshan 5713 5886 8056 7372 10231 9035Baoan 3367 2012 3821 2185 4197 2323Longgang 4602 3337 4919 3543 6340 4820Longhua 4499 2736 4805 2848 6170 4572Yantian 1630 1285 2030 1411 4931 2858Dapeng 274 373 431 420 431 420Pingshan 1163 1107 1231 1134 1231 1134Guangming 1862 909 1881 927 2835 1297

Transit opportunity of 2011

Transit opportunity of 2020

Transit opportunity of 2016

0 40

N

Km

Transit Opportunity0000000 - 25000000002500000001 - 50000000005000000001 - 1000000000010000000001 - 1500000000015000000001 - 2000000000020000000001 - 3000000000030000000001 - 4000000000040000000001 - 80000000000

Figure 3 Maps of transit opportunity in three periods

of the core area Yantian District is a natural scenic touristarea and port area Although it is located in the subcentralarea its public transportation system is not well-developedso the transit opportunities are lower than other subcentersThe rest regions are in the outer suburbs of Shenzhenwhich are far from the core area with a small populationand minimal opportunity The opportunity of all regions isgrowing over time for both groups Transit opportunity ofcore area (especially in Futian) proliferated more than 25from 2011 to 2016 Subcenters and suburbs had a smallergrowth lower than 10 Meanwhile the differences in theabsolute value of transit opportunity between the core areasubcenter area and the suburbs were unusually large and willbe more significant in 2020 It indicates that the residents inthe city center have the most significant benefit from publictransport system

Table 4 shows the average transit opportunity in thewhole city In each measure year transit opportunity of NUVis greater than UVG The transit opportunity of UVG willincrease 45 from 2011 to 2020 and NUV will increase 63at the same time The difference in growth rate is 22When comparing the absolute value of transit opportunitythe difference between the two groups has almost doubledfrom 2011 to 2020 in 2011 NUVrsquos transit opportunity was 3279more than UVGrsquos and the gap will be 6102 in 2020 Fromthe perspective of the city NUV has more public transportadvantages than UVG For subgroups it is clear that thecore area has the highest transit opportunities whether itis UVG or NUV Subgroup 4 (residents who live in theurban village at core area) even has higher opportunities thansubgroup 2 (residents who live in formal urban housing atsubcenter)

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Journal of Advanced Transportation 7

Table 2 Subgroups of NUV and UVG in Shenzhen

Group Subgroup Description Population of 2011 Percent

NUV

1 Residents who live in formal urban housing at core area(Futian Luohu and Nanshan) 4007619 278

2 Residents who live in formal urban housing atsubcenter (Longhua Yantian Longgang and Baoan) 3505463 2369

3 Residents who live in formal urban housing at suburb(Dapeng Pingshan and Guangming)

162948 037

Total 7676030 5186

UVG

4 Residents who live in the urban village at core area(Futian Luohu and Nanshan) 1402041 948

5 Residents who live in the urban village of subcenter(Longhua Yantian Longgang and Baoan)

4724083 3193

6 Residents who live in the urban village at suburb(Dapeng Pingshan and Guangming)

995467 673

Total 7121591 4814

LegendMetro line 2011Metro line 2016Metro line 2020Bus line 2016Core AreaSubcenterSuburb

0 40

Kilometers

N

Figure 2 Maps of the transit network in three periods

until 2020 so it is essential to analyze the impact of publictransport networks on fairness When calculating potentialopportunity of residents our definition of opportunity refersto jobs UPDIS provides a detailed building data of Shenzhenso the floor area of building in employment site is calculatedto represent the opportunity

5 Results and Discussions

51 The Average Transit Opportunity of Two Groups in EachDistrict The minimum unit for calculating transit oppor-tunity is the residential land unit and transit opportunityat the different aggregate levels are calculated by populationweights Figure 3 shows transit opportunity for the differentyears in Shenzhen at the community level These maps

illustrate how varied are the distributions of transit opportu-nity among the region by public bus and metro

We examined the average transit opportunity of twogroups in each district Table 3 shows the average transitopportunity of two groups at the region level Shenzhenrsquos corearea has the most significant opportunity Futian which is thecity center has the highest transit opportunities Luohu is theformer city center in the 1990s and has the second highestaccessibility Nanshan is the critical development area in thefuture with the third highest transit opportunities Thesethree regions are spatially adjacent and possess the mostpublic transport resources which have many subway linesand bus routes Baoan Longhua and Longgang are locatedoutside the core area which is the subcenter of ShenzhenThe transit opportunity in these areas is about one-third

8 Journal of Advanced Transportation

Table 3 Transit opportunity of two groups in each region

Year 2011 2016 2020NUV UVG NUV UVG NUV UVG

Futian 12381 13156 19592 21091 23095 23848Luohu 12477 13041 16143 15106 18823 17035Nanshan 5713 5886 8056 7372 10231 9035Baoan 3367 2012 3821 2185 4197 2323Longgang 4602 3337 4919 3543 6340 4820Longhua 4499 2736 4805 2848 6170 4572Yantian 1630 1285 2030 1411 4931 2858Dapeng 274 373 431 420 431 420Pingshan 1163 1107 1231 1134 1231 1134Guangming 1862 909 1881 927 2835 1297

Transit opportunity of 2011

Transit opportunity of 2020

Transit opportunity of 2016

0 40

N

Km

Transit Opportunity0000000 - 25000000002500000001 - 50000000005000000001 - 1000000000010000000001 - 1500000000015000000001 - 2000000000020000000001 - 3000000000030000000001 - 4000000000040000000001 - 80000000000

Figure 3 Maps of transit opportunity in three periods

of the core area Yantian District is a natural scenic touristarea and port area Although it is located in the subcentralarea its public transportation system is not well-developedso the transit opportunities are lower than other subcentersThe rest regions are in the outer suburbs of Shenzhenwhich are far from the core area with a small populationand minimal opportunity The opportunity of all regions isgrowing over time for both groups Transit opportunity ofcore area (especially in Futian) proliferated more than 25from 2011 to 2016 Subcenters and suburbs had a smallergrowth lower than 10 Meanwhile the differences in theabsolute value of transit opportunity between the core areasubcenter area and the suburbs were unusually large and willbe more significant in 2020 It indicates that the residents inthe city center have the most significant benefit from publictransport system

Table 4 shows the average transit opportunity in thewhole city In each measure year transit opportunity of NUVis greater than UVG The transit opportunity of UVG willincrease 45 from 2011 to 2020 and NUV will increase 63at the same time The difference in growth rate is 22When comparing the absolute value of transit opportunitythe difference between the two groups has almost doubledfrom 2011 to 2020 in 2011 NUVrsquos transit opportunity was 3279more than UVGrsquos and the gap will be 6102 in 2020 Fromthe perspective of the city NUV has more public transportadvantages than UVG For subgroups it is clear that thecore area has the highest transit opportunities whether itis UVG or NUV Subgroup 4 (residents who live in theurban village at core area) even has higher opportunities thansubgroup 2 (residents who live in formal urban housing atsubcenter)

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

8 Journal of Advanced Transportation

Table 3 Transit opportunity of two groups in each region

Year 2011 2016 2020NUV UVG NUV UVG NUV UVG

Futian 12381 13156 19592 21091 23095 23848Luohu 12477 13041 16143 15106 18823 17035Nanshan 5713 5886 8056 7372 10231 9035Baoan 3367 2012 3821 2185 4197 2323Longgang 4602 3337 4919 3543 6340 4820Longhua 4499 2736 4805 2848 6170 4572Yantian 1630 1285 2030 1411 4931 2858Dapeng 274 373 431 420 431 420Pingshan 1163 1107 1231 1134 1231 1134Guangming 1862 909 1881 927 2835 1297

Transit opportunity of 2011

Transit opportunity of 2020

Transit opportunity of 2016

0 40

N

Km

Transit Opportunity0000000 - 25000000002500000001 - 50000000005000000001 - 1000000000010000000001 - 1500000000015000000001 - 2000000000020000000001 - 3000000000030000000001 - 4000000000040000000001 - 80000000000

Figure 3 Maps of transit opportunity in three periods

of the core area Yantian District is a natural scenic touristarea and port area Although it is located in the subcentralarea its public transportation system is not well-developedso the transit opportunities are lower than other subcentersThe rest regions are in the outer suburbs of Shenzhenwhich are far from the core area with a small populationand minimal opportunity The opportunity of all regions isgrowing over time for both groups Transit opportunity ofcore area (especially in Futian) proliferated more than 25from 2011 to 2016 Subcenters and suburbs had a smallergrowth lower than 10 Meanwhile the differences in theabsolute value of transit opportunity between the core areasubcenter area and the suburbs were unusually large and willbe more significant in 2020 It indicates that the residents inthe city center have the most significant benefit from publictransport system

Table 4 shows the average transit opportunity in thewhole city In each measure year transit opportunity of NUVis greater than UVG The transit opportunity of UVG willincrease 45 from 2011 to 2020 and NUV will increase 63at the same time The difference in growth rate is 22When comparing the absolute value of transit opportunitythe difference between the two groups has almost doubledfrom 2011 to 2020 in 2011 NUVrsquos transit opportunity was 3279more than UVGrsquos and the gap will be 6102 in 2020 Fromthe perspective of the city NUV has more public transportadvantages than UVG For subgroups it is clear that thecore area has the highest transit opportunities whether itis UVG or NUV Subgroup 4 (residents who live in theurban village at core area) even has higher opportunities thansubgroup 2 (residents who live in formal urban housing atsubcenter)

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Journal of Advanced Transportation 9

Table 4 The average transit opportunity of two groups in the whole city

Year 2011 2016 2020Group NUV 7307 9959 11972

Subgroup1 10597 15367 182492 3819 4172 52553 1457 1530 2090

Group UVG 4028 4846 5870

Subgroup4 10825 14397 164805 2662 2835 37266 943 987 1102

Table 5 Decomposition of the Gini coefficient between UVG and NUV

Year 2011 2016 2020Total Gini G 05725 05916 05736

Gini within Gnuk 05049 05160 04920Gukg 06267 06480 06385

Gini between Gukg VS Gnuk 06026 06303 06140

ContributionGw 4814 4763 4738Gb 2504 2884 2945Gt 2681 2353 2317

Gw contribution of the Gini inequality indexes within subpopulations to the total Gini ratioGb contribution of the Gini inequality ratios between subpopulations to the total Gini ratioGt contribution of the transit opportunity intensity of transvariation between subpopulations to the total Gini ratio

52 Changes of Transit Opportunity Distribution

521 Decomposition of Transit Opportunity Gini Coefficientbetween NUV and UVG To analyze the horizontal equityof public transport we calculated and decomposed the Ginicoefficient using the transit opportunity of 7625 residentialunits the total populations of Shenzhen was divided into thefollowing subgroups UVG and NUV Table 5 presents thedecomposition of the Gini coefficients estimated for the twogroupsThe total Gini119866 reflects the equity situation of transitopportunity distribution in all populations of ShenzhenFrom 2011 to 2020 its value changed from 05725 to 05736Although the changes are not significant we still can seethe trends of transit opportunity equity Compared withthe public transport network in 2011 there were three newsubway lines added to the public transport network in 2016(in Figure 2 green lines) These lines are mainly locatedin the core area and it caused an increase of inequity forall populations In 2020 the subway network will expandfrom the core area to the subcenter area (in Figure 2 bluelines) The total Gini coefficient will be reduced that is thedistribution of transit opportunity in 2020 will be more eventhan in 2016 For horizontal equity of the total populationthe gap in transit opportunity between the overall populationincreases first and then decreases

As for within group inequity Gukg is the equity indexof transit opportunity distribution for UVG population andGnuk is the equity index for NUV population The resultshows that in each measure year Gnuk is less than Gukg sothe transit opportunity distribution of NUV ismore equitablethan UVG and the gap of Gini coefficients between UVG

and NUV is significant from 01218 in 2011 to 01465 in 2020The impact of changes in the public transport system ontransportation equity is consistent for both NUV and UVGIn 2016 the Gini coefficient increased in both groups andthe Gini coefficient of both groups will decrease by 2020However in 2002 the NUVrsquos distribution of opportunity ismore equitable than the distribution in 2011 UVGs will notreturn to the level in 2011 which means that Gnkg (2020)gt Gnkg (2011)

The between-group inequity (Gukg VS Gnuk) results showthat the public transport network in 2016 resulted in themost considerable inequality between the two groups Withthe expansion of the subway network from the core area toother areas in 2020 the inequality between the two groupswill decrease The analysis of inequality contribution showsthat the within groups inequality (Gw) has contributed themost to the total inequity in three periods The contribu-tion of inequality between groups (Gb) is growing so thedevelopment of public transport has led to growing inequalitybetween the two groups

522 Decomposition of Transit Opportunity Gini Coefficientbetween Subgroups We know that within groups inequalityhas the most significant impact on overall inequality inTable 5 so we decomposed the Gini coefficient for eachgroup and explored about the influence of spatial location ofresidence on equity Table 6 presents the decomposition oftransit opportunity Gini coefficient Gnuk between subgroupsof NUVG119894means the equity index of transit opportunity dis-tribution of subgroup i The result shows that most equitable

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

10 Journal of Advanced Transportation

Table 6 Decomposition of the Gini coefficient of NUV between subgroups

Year 2011 2016 2020Gnuk 05049 05159 04920

Gini within NUVG1 04060 03685 03398G2 05045 05192 05131G3 05097 05101 05458

ContributionGw(nuk) 426 3875 3819Gb(nuk) 4732 5568 5636Gt(nuk) 108 557 545

Table 7 Decomposition of the Gini coefficient of UVG between subgroups

Year 2011 2016 2020Gukg 06266 06480 06385

Gini within UVGG4 04333 03748 03567G5 05441 05525 05760G6 05424 05463 05572

ContributionGw(ukg) 3276 2885 3153Gb(ukg) 5951 6566 6246ampGt(ukg) 773 549 601

Table 8 The Gini coefficients and transit opportunities of all subgroups

Year2011 2016 2020

subgroup Gi ATO subgroup Gi ATO subgroup Gi ATO1 0406 10597 1 03685 15367 1 03398 182494 04333 10825 4 03748 14397 4 03567 164802 05045 3819 3 05101 1530 2 05131 52553 05097 1457 2 05192 4172 3 05458 20906 05424 943 6 05463 987 6 05572 11025 05441 2662 5 05525 2835 5 0576 3726Gi transit opportunity Gini coefficient of subgroup iATO average transit opportunity

distribution of transit opportunity inNUV is subgroup 1 in allthree years Compared with G2 and G3 G1 is the smallest andis consistently decreasing in three periods and its decreaseis also the largest G2 and G3 are almost equal in 2011 and2016 but the value of G3 will be larger in 2020 The analysisof inequality contribution of Gnuk shows that the betweengroups inequality (Gb(nuk)) has contributed the most to thetotal inequity in three periods of NUV

Table 7 presents the decomposition of transit opportunityGini coefficient Gukg between subgroups of UVG For UVGTable 7 shows that G4 is significantly different from G5 andG6 and themost equitable distribution of transit opportunityis subgroup 4 From the changes in the Gini coefficientthe inequality of all UVG population has increased andthe opportunity distribution gap of UVG population in corearea tends to be smaller Subgroup 5 is the most unfairdistribution of opportunities in UVG G6 is slightly smallerthan G5 The construction of subway infrastructure from2011 to 2020 will have a negative impact on the distributionof transit opportunity in subcenters and suburbs From the

analysis of the contribution to overall inequality of UVG theinequality among the groups Gb(ukg) contributes the mostaccounting for about 60 This shows that the unfairnessbetween regions mainly causes the overall inequality ofUVG

Table 8 summarizes the Gini coefficients and transitopportunities of all subgroups The order of the subgroupsin the table is arranged according to the value of theGini coefficient For each region (core area subcentersand suburb) the Gini coefficient of NUV is always smallerthan UVG in all three years and transit opportunity ofNUV is larger than of UVG Residents of NUV in the coreareas have the highest opportunities and the most equitabledistribution Residents of UVG in the core areas have thesecond highest opportunities and equitable distributionFor subcenters and suburbs the opportunity gap betweenthe two regions is significant but the difference in equitydistribution is not apparent With the development of urbanpublic transportation systems the absolute value of residentsrsquotransit opportunities is improved but it may lead to unfair

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Journal of Advanced Transportation 11

2 3 4 5 6 7 8 9 10 11 120

02

04

06

08

Kern

el d

ensit

y

Distribution of transit opportunity of NUV in three periods

201120162020

0

02

04

06

08

2 3 4 5 6 7 8 9 10 11 12LOG(transit opportunity)

LOG(transit opportunity)

Kern

el d

ensit

y

Distribution of transit opportunity of UVG in three periods

201120162020

Figure 4 Distribution of transit opportunity probability of NUV and UVG population

distribution From 2011 to 2020 transit opportunities ofsubgroup 5 and subgroup 6 will rise but equity performancewill fall Spatial location and low-income both have impacton equity but spatial location plays a significant role in thedifference and opportunities distributions between groupsespecially for the UVG

53 Changes of Transit Opportunity Probability DistributionBased on the transit opportunity of 7825 settlements we usedkernel density estimation to get the probability of transitopportunity and plotted the probability density distributionmap

531 The Comparisons of Probability Density Maps betweenUVG and NUV Figure 4 shows the distribution of transitopportunity probability of NUV and UVG The horizontalaxis is the value of transit opportunity Since the range ofopportunity in different residential areas is too large a LOGconversion is performed The vertical axis is the probabilityof transit opportunity

The top of Figure 4 is transit opportunity probability ofNUV From 2011 to 2020 there is a slight right movementof the curve which means the value of the opportunitywill increase The density curve in 2016 is quite different

from in 2011 and it shows that the subway line in 2016has a significant influence on the opportunity distributionof NUV population In 2011 the probability of opportunityvalue ldquo8rdquo and opportunity value ldquo9rdquo was not much differentConsidering vertical equity an ideal situation is that thepublic transport system prioritizes raising the odds of loweropportunity population and reduces the probability of loweropportunity The person with the opportunity ldquo8rdquo shouldobtain the priority of improvement In fact the probabilityof ldquo9rdquo was reduced more than the probability of ldquo8rdquo in2016 which means that the public transport system is moreinclined to improve the population with high initial oppor-tunity and make them higher Therefore the distribution ofunfairness increased The probability density curve shape in2016 is very similar to that in 2020 and it indicates thatthe subway line in 2020 will have a little influence on theopportunity distribution of NUV population The tendencyof concentration will be more obvious and further reducethe probability of around ldquo8rdquo in 2020 and distribution ofunfairness will decrease The bottom of Figure 4 showsthe distribution of transit opportunity of UVG populationThere is no significant change in the density curve shapecompared with NUV and it means that from 2011 to 2020the improvement of the public transportation system has asmaller impact than NUV It has the same phenomenon with

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

12 Journal of Advanced Transportation

001020304

Distribution of transit opportunity of NUV and UVG in 2011

2 3 4 5 6 7 8 9 10 11 12

NUVUVG

2 3 4 5 6 7 8 9 10 11 120

01020304

Kern

el d

ensit

y Distribution of transit opportunity of NUV and UVG in 2016

NUVUVG

01020304

Distribution of transit opportunity of NUV and UVG in 2020

2 3 4 5 6 7 8 9 10 11 120

LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

NUVUVG

Figure 5 Comparisons of the probability distribution of the two groups in the whole city

NUV which has a priority to improve the high opportunitypopulation Besides the curves have three peaks in 2016 andin 2020 which shows that multipolarity is more severe andnoticeable

Figure 5 shows the comparisons of the probability dis-tribution of the two groups at the whole city level We canobserve that the opportunity value ldquo8rdquo is a dividing point in2011 and 2016 For transit opportunity which value is less than8 the probability of UVG is higher than NUV This meansthat the UVG population is more likely to have low publictransport opportunities For transit opportunity which valueis larger than 8 the probability of UVG is less than NUVso the NUV population is more likely to have high publictransport opportunities In 2020 the dividing point of the twogroup is 85 and this means that the opportunities of bothgroupswill be improvedThe comparison of the two groups inthe three periods illustrates Shenzhen public transportationsystem is very favorable to NUV in all three periods Theprobability gap of high transit opportunity between the twogroups is increasing in different periods this also shows thatNUV benefits more than UVG from the improvement of theurban transport system

532 The Comparisons of Probability Density Maps betweenSubgroups Figure 6 is the comparisons of the probabilitydistribution of the two groups in core areasThe distributionsof the two groups are very similar The width of the wave inthe distribution curve of the two groups becomes narrowerover time indicating that the difference of transit opportunitybetween population is smaller Figure 7 is comparisons ofthe probability distribution of the two groups in subcenterCompared with Figure 6 the difference in probability dis-tribution curves between the two groups is noticeable andthe width of the wave in the distribution curve is wider thanthe core areas From 2011 to 2020 the probability gap of hightransit opportunity between UVG and NUV is getting largerThe distributions of transit opportunities between the twogroups of suburbs have the same characteristics as that of thesubcenter in Figure 8 including broader wave and significantdistribution differences between UVG and NUV The threefigures show that there is a difference in the distributionof opportunities between different regions and there arealso differences between the two groups in the same regionSpatial location and type of group have an impact on accessto transit opportunities

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Journal of Advanced Transportation 13

0 2 4 6 8 10 120

05

1Distribution of transit opportunity of core area in 2011

NUVUVG

0 2 4 6 8 10 120

05

1

Kern

el d

ensit

y Distribution of transit opportunity of core area in 2016

NUVUVG

0

05

1

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of core area in 2020

NUVUVG

Figure 6 Comparisons of the probability distribution of the two groups in core areas

6 Conclusions

Decade-long changes in disparity and distribution of transitopportunity gave us a clear picture and the results illustratedthe importance of examining transportation equity over along period (1) For the absolute value of the opportunityShenzhenrsquos core area has the most significant opportunityand urban village populations do have fewer transporta-tion opportunities than nonurban villages People in thedifferent regions and groups inevitably do not have equaltransit opportunity and this is not necessarily problematicTransport policy is fair if it distributes transport investmentsand services in ways that reduce inequality of opportu-nity With the development of Shenzhen public transportinfrastructures although all populations are benefiting fromincreasing transit opportunity transit opportunity of NUVis greater than UVG Their gap is widening in each mea-sure year and it is necessary to limit the highest levels ofaccessibility of social groups when a marginal improvementof accessibility at the upper levels would harm those groupsat the bottom (2) For the vertical equity the public trans-port system is more inclined to improve the populationwith high initial opportunity and make them higher The

improvement of the public transportation system has asmaller impact than NUV in all three periods and poli-cies should prioritize disadvantaged groups (3) From thehorizontal equity of transit opportunity distribution forall population the development of public transport in dif-ferent periods in Shenzhen first exacerbated unfair distri-bution of transit opportunity and then increased fairnessThe development of transit also has a different impacton the different group of people In each measure yearthe distribution of NUV is more equitable than UVGand the gap of Gini coefficient between the two groupsis significant and is widening (4) Peoplersquos social statusand spatial location are both factors that contribute tothe total inequity The spatial location has more impacton equity and low-income people do not necessarily havetraffic disadvantages Low-income people living in thecore area of Shenzhen such as Futian and Luohu havehigh transit opportunity and a more equitable distribu-tion

Our research is beneficial for providing informationto adjust the planning of future Metro routes and urbandevelopment strategies in Shenzhen Since social status andspatial location are factors of impact equity Shenzhen should

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

14 Journal of Advanced Transportation

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of subcenter in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of subcenter in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of subcenter in 2020

NUVUVG

Figure 7 Comparisons of the probability distribution of the two groups in subcenter

strengthen public transport services in subcentral areasespecially Baoan and Longgang The two regions have a largepopulation for both NUV and UVG and improvement ofpublic transport in these areas will be the most effectiveway for improving public transport accessibility and fairdistribution Our study also has some limitations Firstwe used the same bus network for the 10-year analysisperiod and this operation would cause deviations in thecalculation results of transit opportunities So the analysisof equity ignored the changes in fairness brought about bythe improvement of bus service level Second transportationequity not only is an infrastructure issue but also involvesland use planning as well The transit opportunity of aresident is related to the public transport system and to thedistribution and size of the opportunities The distributionof opportunities is related to urban planning especiallyland use This study does not explore the impact of landuse on transportation equity Third it is not only thedistribution of access to destinations that matter but alsoin some cases the absolute level of access for those whoare worse We only discussed the changes of the residentsrsquoopportunities and the changes of disparity and distributionof transit opportunity and we do not discuss whethertransit opportunities meet the needs of different groups of

people and their satisfaction of transportation opportunitiesIn the future research those are our next research direc-tion

Data Availability

The population distribution data employment distributiondata and transit data of Shenzhen used to support thefindings of this study were supplied by Shenzhen UrbanPlanning Bureau (SUPB) and theUrban Planning andDesignInstitute of Shenzhen (UPDIS) under license and so cannot bemade freely available Requests for access to these data shouldbe made to Qingfeng ZHOU zhouqingfenghiteducn

Conflicts of Interest

The authors declarethat they have no conflicts of interest

Acknowledgments

The authors gratefully acknowledge financial support fromthe China Scholarship Council The authors gratefullyacknowledge support from Program on Chinese Cities of

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Journal of Advanced Transportation 15

0 2 4 6 8 10 120

01

02

03

04Distribution of transit opportunity of suburb in 2011

NUVUVG

0 2 4 6 8 10 120

01

02

03

04

Kern

el d

ensit

y Distribution of transit opportunity of suburb in 2016

NUVUVG

0

01

02

03

04

0 2 4 6 8 10 12LOG(transit opportunity)

LOG(transit opportunity)

LOG(transit opportunity)

Distribution of transit opportunity of suburb in 2020

NUVUVG

Figure 8 Comparisons of the probability distribution of the two groups in the suburb

Center for Urban and regional Studies University of NorthCarolina at Chapel Hill

References

[1] J Jain T Line and G Lyons ldquoA troublesome transport chal-lenge Working round the school runrdquo Journal of TransportGeography vol 19 no 6 pp 1608ndash1615 2011

[2] G Currie and A Delbosc ldquoModelling the social and psycholog-ical impacts of transport disadvantagerdquo Transportation vol 37no 6 pp 953ndash966 2010

[3] A Lubin and D Deka ldquoRole of public transportation as jobaccess mode lessons fromfbfbsurvey of people with disabilitiesin New Jerseyrdquo Transportation Research Record vol 2277 pp90ndash97 2012

[4] A Matas J-L Raymond and J-L Roig ldquoJob accessibility andfemale employment probability The cases of Barcelona andMadridrdquo Urban Studies vol 47 no 4 pp 769ndash787 2010

[5] A Cebollada ldquoMobility and labour market exclusion in theBarcelona Metropolitan Regionrdquo Journal of Transport Geogra-phy vol 17 no 3 pp 226ndash233 2009

[6] K Martens ldquoSubstance precedes methodology On cost-benefitanalysis and equityrdquo Transportation vol 38 no 6 pp 959ndash9742011

[7] K Martens A Golub and G Robinson ldquoA justice-theoreticapproach to the distribution of transportation benefits Implica-tions for transportation planning practice in the United StatesrdquoTransportation Research Part A Policy and Practice vol 46 no4 pp 684ndash695 2012

[8] N Thomopoulos and S Grant-Muller ldquoIncorporating equityas part of the wider impacts in transport infrastructure assess-mentAn application of the SUMINI approachrdquoTransportationvol 40 no 2 pp 315ndash345 2013

[9] G Currie ldquoQuantifying spatial gaps in public transport supplybased on social needsrdquo Journal of Transport Geography vol 18no 1 pp 31ndash41 2010

[10] B Epstein andM Givoni ldquoAnalyzing the gap between the QOSdemanded by PT users and QOS supplied by service operatorsrdquoTransportation Research Part A Policy and Practice vol 94 pp622ndash637 2016

[11] C Jaramillo C Lizarraga and A L Grindlay ldquoSpatial disparityin transport social needs and public transport provision inSantiago de Cali (Colombia)rdquo Journal of Transport Geographyvol 24 pp 340ndash357 2012

[12] E M Ferguson J Duthie A Unnikrishnan and S T WallerldquoIncorporating equity into the transit frequency-setting prob-lemrdquo Transportation Research Part A Policy and Practice vol46 no 1 pp 190ndash199 2012

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

16 Journal of Advanced Transportation

[13] A Monzon E Ortega and E Lopez ldquoEfficiency and spatialequity impacts of high-speed rail extensions in urban areasrdquoCities vol 30 no 1 pp 18ndash30 2013

[14] J P Bocarejo S and D R Oviedo H ldquoTransport accessibilityand social inequities a tool for identification of mobility needsand evaluation of transport investmentsrdquo Journal of TransportGeography vol 24 pp 142ndash154 2012

[15] B van Wee and K Geurs ldquoDiscussing equity and social exclu-sion in accessibility evaluationsrdquo European Journal of Transportand Infrastructure Research vol 11 no 4 pp 350ndash367 2011

[16] D Boucher and P J Kelly Social justice from Hume to WalzerPsychology Press 1998

[17] T Litman ldquoEvaluating transportation equityrdquo World TransportPolicy amp Practice vol 8 no 2 pp 50ndash65 2002

[18] F Di Ciommo and Y Shiftan ldquoTransport equity analysisrdquoTransport Reviews vol 37 no 2 pp 139ndash151 2017

[19] C Falavigna and D Hernandez ldquoAssessing inequalities onpublic transport affordability in two latin American citiesMontevideo (Uruguay) and Cordoba (Argentina)rdquo TransportPolicy vol 45 pp 145ndash155 2016

[20] A T Murray ldquoStrategic analysis of public transport coveragerdquoSocio-Economic Planning Sciences vol 35 no 3 pp 175ndash1882001

[21] A Delbosc andG Currie ldquoUsing Lorenz curves to assess publictransport equityrdquo Journal of Transport Geography vol 19 no 6pp 1252ndash1259 2011

[22] K Fransen T Neutens S Farber P De Maeyer G Deruyterand F Witlox ldquoIdentifying public transport gaps using time-dependent accessibility levelsrdquo Journal of Transport Geographyvol 48 pp 176ndash187 2015

[23] T Saghapour S Moridpour and R G Thompson ldquoPublictransport accessibility in metropolitan areas A new approachincorporating population densityrdquo Journal of Transport Geogra-phy vol 54 pp 273ndash285 2016

[24] E C Delmelle and I Casas ldquoEvaluating the spatial equity ofbus rapid transit-based accessibility patterns in a developingcountry The case of Cali Colombiardquo Transport Policy vol 20pp 36ndash46 2012

[25] J Grengs ldquoNonwork Accessibility as a Social Equity IndicatorrdquoInternational Journal of Sustainable Transportation vol 9 no 1pp 1ndash14 2015

[26] A Paez D M Scott and C Morency ldquoMeasuring accessibilityPositive and normative implementations of various accessibilityindicatorsrdquo Journal of Transport Geography vol 25 pp 141ndash1532012

[27] A El-GeneidyD Levinson EDiab G Boisjoly DVerbich andC Loong ldquoThe cost of equity Assessing transit accessibility andsocial disparity using total travel costrdquo Transportation ResearchPart A Policy and Practice vol 91 pp 302ndash316 2016

[28] M Niehaus P Galilea and R Hurtubia ldquoAccessibility andequity An approach for wider transport project assessment inChilerdquo Research in Transportation Economics vol 59 pp 412ndash422 2016

[29] C Dagum ldquoA new approach to the decomposition of the Giniincome inequality ratiordquo in Income Inequality Poverty andEconomic Welfare pp 47ndash63 Physica-Verlag HD 1998

[30] M Rosenblatt ldquoRemarks on some nonparametric estimates of adensity functionrdquo Annals of Mathematical Statistics vol 27 pp832ndash837 1956

[31] J Zacharias and Y Tang ldquoRestructuring and repositioningShenzhen Chinarsquos new mega cityrdquo Progress in Planning vol 73no 4 pp 209ndash249 2010

[32] Y P Wang Y Wang and J Wu ldquoHousing migrant workers inrapidly urbanizing regions A study of the Chinese model inShenzhenrdquoHousing Studies vol 25 no 1 pp 83ndash100 2010

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom