GEOGRAPHY

6000 Broken Sound Parkway, NW Suite 300, Boca Raton, FL 33487711 Third Avenue New York, NY 100172 Park Square, Milton Park Abingdon, Oxon OX14 4RN, UK

an informa business

w w w . c r c p r e s s . c o mFahui Wang

w w w . c r c p r e s s . c o m

K19056

Quantitative Methods and Socio-Economic Applications in

GIS

Qu

an

titative

Me

tho

ds a

nd

Soc

io-Ec

on

om

ic A

pp

lica

tion

s in G

IS

SecondEdition

Wang Second Edition

“This is a thoroughly revised edition of a highly useful book on quantitative and GIS methods. … provides useful discussion of important methodological issues such as spatial autocorrelation and the modifiable areal unit problem. … an excellent book for students and social scientists interested in various social and economic phenomena with spatial manifestations.”—Mei-Po Kwan, University of Illinois at Urbana-Champaign

“… an excellent book with clear, meaningful, and well-designed examples. It’s one of the most useful books on my shelf!”—Sara McLafferty, University of Illinois at Urbana-Champaign

See What’s New in the Second Edition:

• All project instructions are in ArcGIS 10.2 using geodatabase datasets

• New chapters on regionalization methods and Monte Carlo simulation

• Popular tasks automated as a convenient toolkit: Huff Model, 2SFCA accessibility measure, regionalization, Garin–Lowry model, and Monte Carlo–based spatial simulation

• Advanced tasks now implemented in user-friendly programs or ArcGIS: centrality indices, wasteful commuting measure, p-median problem, and traffic simulation

The second edition of a bestseller, Quantitative Methods and Socio-Economic Applications in GIS (previously titled Quantitative Methods and Applications in GIS) details applications of quantitative methods in social science, planning, and public policy with a focus on spatial perspectives. The book integrates GIS and quantitative (computational) methods and demonstrates them in various policy-relevant socio-economic applications with step-by-step instructions and datasets. The book demonstrates the diversity of issues where GIS can be used to enhance the studies related to socio-economic issues and public policy.

Second Edition

Quantitative Methods and Socio-Economic Applications in

GIS

vii

ContentsList of Figures ..........................................................................................................xvList of Tables ...........................................................................................................xixForeword .................................................................................................................xxiPreface.................................................................................................................. xxiiiAuthor ..................................................................................................................xxviiList of Major GIS Datasets and Program Files .....................................................xxixList of Quick Reference for Spatial Analysis Tasks .............................................xxxi

SECTION I GIS and Basic Spatial Analysis Tasks

Chapter 1 Getting Started with ArcGIS: Data Management and Basic Spatial Analysis Tools ..........................................................................3

1.1 Spatial and Attribute Data Management in ArcGIS .................31.1.1 Map Projections and Spatial Data Models ...................41.1.2 Attribute Data Management and Attribute Join ...........5

1.2 Spatial Analysis Tools in ArcGIS: Queries, Spatial Joins, and Map Overlays ......................................................................7

1.3 Case Study 1: Mapping and Analyzing Population Density Pattern in Baton Rouge, Louisiana ............................ 101.3.1 Part 1: Mapping the Population Density Pattern

Across Census Tracts ................................................. 101.3.2 Part 2: Analyzing the Population Density Pattern

Across Concentric Rings ............................................ 161.4 Summary .................................................................................23Appendix 1: Identifying Contiguous Polygons by Spatial

Analysis Tools .........................................................................23

Chapter 2 Measuring Distance and Time ...........................................................27

2.1 Measures of Distance ..............................................................272.2 Computing Network Distance and Time .................................292.3 The Distance Decay Rule ........................................................ 322.4 Case Study 2: Computing Distances and Travel Time to

Public Hospitals in Louisiana .................................................. 332.4.1 Part 1: Measuring Euclidean and Manhattan

Distances ....................................................................342.4.2 Part 2: Measuring Travel Time .................................. 38

2.5 Summary ................................................................................. 41

K19056_C000toc.indd 7 9/26/2014 10:44:23 PM

viii Contents

Appendix 2A: The Valued Graph Approach to the Shortest Route Problem ......................................................................... 41

Appendix 2B: Estimating Travel Time Matrix by Google Maps API .... 42

Chapter 3 Spatial Smoothing and Spatial Interpolation ..................................... 47

3.1 Spatial Smoothing ................................................................... 473.1.1 Floating Catchment Area (FCA) Method ..................483.1.2 Kernel Density Estimation ......................................... 49

3.2 Point-Based Spatial Interpolation ............................................503.2.1 Global Interpolation Methods ....................................503.2.2 Local Interpolation Methods ...................................... 51

3.3 Case Study 3A: Mapping Place Names in Guangxi, China ...... 533.3.1 Part 1: Spatial Smoothing by the Floating

Catchment Area Method ............................................ 533.3.2 Part 2: Spatial Interpolation by Various Methods ...... 56

3.4 Area-Based Spatial Interpolation ............................................ 593.5 Case Study 3B: Area-Based Interpolations of Population

in Baton Rouge, Louisiana ......................................................603.5.1 Part 1. Using the Areal Weighting Interpolation

to Transform Data from Census Tracts to School Districts in 2010 .........................................................60

3.5.2 Part 2. Using the Target-Density Weighting (TDW) Interpolation to Interpolate Data from Census Tracts in 2010 to Census Tracts in 2000 ........ 61

3.6 Summary .................................................................................64Appendix 3A: Empirical Bayes Estimation for Spatial Smoothing .......64Appendix 3B: The Network Hierarchical Weighting Method for

Areal Interpolation ...................................................................65

SECTION II Basic Quantitative Methods and Applications

Chapter 4 GIS-Based Trade Area Analysis and Application in Business Geography ...........................................................................69

4.1 Basic Methods for Trade Area Analysis .................................. 704.1.1 Analog Method and Regression Models .................... 704.1.2 Proximal Area Method ............................................... 70

4.2 Gravity Models for Delineating Trade Areas .......................... 724.2.1 Reilly’s Law ................................................................ 724.2.2 Huff Model ................................................................. 734.2.3 Link between Reilly’s Law and Huff Model .............. 744.2.4 Extensions of the Huff Model .................................... 75

K19056_C000toc.indd 8 9/26/2014 10:44:23 PM

ixContents

4.3 Case Study 4A: Defining Fan Bases of Cubs and White Sox in Chicago Region ................................................. 764.3.1 Part 1. Defining Fan Base Areas by the Proximal

Area Method............................................................... 784.3.2 Part 2. Defining Fan Base Areas and Mapping

Probability Surface by Huff Model ............................ 794.3.3 Discussion................................................................... 81

4.4 Case Study 4B: Estimating Trade Areas of Public Hospitals in Louisiana ............................................................. 824.4.1 Part 1. Defining Hospital Service Areas by the

Proximal Area Method ............................................... 824.4.2 Part 2. Defining Hospital Service Areas by Huff

Model .......................................................................... 834.5 Concluding Remarks ...............................................................87Appendix 4A: Economic Foundation of the Gravity Model ..............88Appendix 4B: A Toolkit for Implementing the Huff Model ..............90

Chapter 5 GIS-Based Measures of Spatial Accessibility and Application in Examining Health Care Access .........................................................93

5.1 Issues on Accessibility .............................................................935.2 The Floating Catchment Area Methods ..................................95

5.2.1 Earlier Versions of Floating Catchment Area (FCA) Method ............................................................95

5.2.2 Two-Step Floating Catchment Area (2SFCA) Method .......................................................................96

5.3 The Gravity-Based and Generalized 2SFCA Models .............985.3.1 Gravity-Based Accessibility Index .............................985.3.2 Comparison of the 2SFCA and Gravity-Based

Methods ......................................................................995.3.3 The Generalized 2SFCA Model ............................... 100

5.4 Case Study 5: Measuring Spatial Accessibility to Primary Care Physicians in Chicago Region ........................ 1015.4.1 Part 1. Implementing the 2SFCA Method ................ 1025.4.2 Part 2. Implementing the Gravity-Based

Accessibility Model .................................................. 1055.4.3 Discussion................................................................. 108

5.5 Concluding Comments .......................................................... 108Appendix 5A: A Property of Accessibility Measures ...................... 110Appendix 5B: A Toolkit of Automated Spatial Accessibility

Measures ................................................................................ 112

Chapter 6 Function Fittings by Regressions and Application in Analyzing Urban Density Patterns .................................................................... 115

6.1 The Density Function Approach to Urban and Regional Structures ............................................................................... 115

K19056_C000toc.indd 9 9/26/2014 10:44:23 PM

x Contents

6.1.1 Urban Density Functions .......................................... 1156.1.2 Regional Density Functions ..................................... 117

6.2 Function Fittings for Monocentric Models ............................ 1186.2.1 Four Simple Bivariate Functions .............................. 1186.2.2 Other Monocentric Functions .................................. 1206.2.3 GIS and Regression Implementations ...................... 121

6.3 Nonlinear and Weighted Regressions in Function Fittings ... 1226.4 Function Fittings for Polycentric Models .............................. 126

6.4.1 Polycentric Assumptions and Corresponding Functions .................................................................. 126

6.4.2 GIS and Regression Implementations ...................... 1286.5 Case Study 6: Analyzing Urban Density Patterns in

Chicago Urban Area .............................................................. 1296.5.1 Part 1: Function Fittings for Monocentric Models

at the Census Tract Level ......................................... 1306.5.2 Part 2: Function Fittings for Polycentric Models

at the Census Tract Level ......................................... 1336.5.3 Part 3: Function Fittings for Monocentric Models

at the Township Level............................................... 1346.6 Discussions and Summary .................................................... 136Appendix 6A: Deriving Urban Density Functions ........................... 137Appendix 6B: Centrality Measures and Association with Urban

Densities ................................................................................ 139Appendix 6C: OLS Regression for a Linear Bivariate Model ......... 140

Chapter 7 Principal Components, Factor and Cluster Analyses, and Application in Social Area Analysis ................................................ 143

7.1 Principal Components Analysis ............................................ 1447.2 Factor Analysis ...................................................................... 1457.3 Cluster Analysis ..................................................................... 1497.4 Social Area Analysis ............................................................. 1517.5 Case Study 7: Social Area Analysis in Beijing ..................... 1537.6 Discussions and Summary .................................................... 159Appendix 7: Discriminant Function Analysis .................................. 162

Chapter 8 Spatial Statistics and Applications ................................................... 163

8.1 The Centrographic Measures ................................................ 1648.2 Case Study 8A: Measuring Geographic Distributions of

Racial–Ethnic Groups in Chicago Urban Area ..................... 1668.3 Spatial Cluster Analysis Based on Feature Locations ........... 168

8.3.1 Tests for Global Clustering Based on Feature Locations .................................................................. 168

8.3.2 Tests for Local Clusters Based on Feature Locations .................................................................. 168

K19056_C000toc.indd 10 9/26/2014 10:44:23 PM

xiContents

8.4 Case Study 8B: Spatial Cluster Analysis of Place Names in Guangxi, China ................................................................. 170

8.5 Spatial Cluster Analysis Based on Feature Values ................ 1728.5.1 Defining Spatial Weights .......................................... 1728.5.2 Tests for Global Clustering Based on Feature

Values ....................................................................... 1738.5.3 Tests for Local Clusters Based on Feature Values ... 175

8.6 Spatial Regression ................................................................. 1768.6.1 Spatial Lag Model and Spatial Error Model ............ 1768.6.2 Geographically Weighted Regression ...................... 178

8.7 Case Study 8C: Spatial Cluster and Regression Analyses of Homicide Patterns in Chicago........................................... 1788.7.1 Part 1: Spatial Cluster Analysis of Homicide Rates . 1808.7.2 Part 2: Regression Analysis of Homicide Patterns ... 182

8.8 Summary ............................................................................... 187Appendix 8: Spatial Filtering Methods for Regression Analysis ..... 190

SECTION III Advanced Quantitative Methods and Applications

Chapter 9 Regionalization Methods and Application in Analysis of Cancer Data .................................................................................................. 193

9.1 The Small Population Problem and Regionalization ............ 1939.2 The Spatial Order and the Modified Scale–Space

Clustering (MSSC) Methods ................................................. 1969.3 The REDCAP Method .......................................................... 1999.4 Case Study 9: Constructing Geographical Areas for

Analysis of Late-Stage Breast Cancer Risks in the Chicago Region ..................................................................... 201

9.5 Summary ...............................................................................209Appendix 9A: The Poisson-Based Regression Analysis .................. 212Appendix 9B: A Toolkit of the Mixed-Level Regionalization

Method ................................................................................... 213

Chapter 10 System of Linear Equations and Application of Garin–Lowry Model in Simulating Urban Population and Employment Patterns . 217

10.1 System of Linear Equations................................................... 21710.2 The Garin–Lowry Model ...................................................... 219

10.2.1 Basic versus Nonbasic Economic Activities ............ 21910.2.2 The Model’s Formulation .........................................22010.2.3 An Illustrative Example ........................................... 222

10.3 Case Study 10: Simulating Population and Service Employment Distributions in a Hypothetical City ................ 223

K19056_C000toc.indd 11 9/26/2014 10:44:23 PM

xii Contents

10.4 Discussion and Summary ...................................................... 229Appendix 10A: The Input–Output Model ........................................230Appendix 10B: Solving a System of Nonlinear Equations .............. 231Appendix 10C: A Toolkit for Calibrating the Garin– Lowry Model 233Appendix 10D: Cellular Automata (CA) for Urban Land Use

Modeling ................................................................................ 233

Chapter 11 Linear Programming and Applications in Examining Wasteful Commuting and Allocating Health Care Providers ......................... 237

11.1 Linear Programming and the Simplex Algorithm ................ 23811.1.1 The LP Standard Form ............................................. 23811.1.2 The Simplex Algorithm ............................................ 238

11.2 Case Study 11A: Measuring Wasteful Commuting in Columbus, Ohio ................................................................. 24111.2.1 The Issue of Wasteful Commuting and Model

Formulation .............................................................. 24111.2.2 Data Preparation in ArcGIS .....................................24211.2.3 Measuring Wasteful Commuting in an R program ..245

11.3 Integer Programming and Location–Allocation Problems ... 24711.3.1 General Forms and Solutions for Integer

Programming ........................................................... 24711.3.2 Location–Allocation Problems.................................248

11.4 Case Study 11B: Allocating Health Care Providers in Baton Rouge, Louisiana ......................................................... 251

11.5 SUMMARY ..........................................................................254Appendix 11A: Hamilton’s Model on Wasteful Commuting ...........254Appendix 11B: Coding Linear Programming in SAS ......................256Appendix 11C: A Programming Approach to Minimal Disparity

in Accessibility ...................................................................... 257

Chapter 12 Monte Carlo Method and Its Application in Urban Traffic Simulation ........................................................................................ 259

12.1 Monte Carlo Simulation Method ...........................................26012.1.1 Introduction to Monte Carlo Simulation ..................26012.1.2 Monte Carlo Applications in Spatial Analysis .........260

12.2 Travel Demand Modeling ...................................................... 26212.3 Examples of Monte Carlo-Based Spatial Simulation ............26412.4 Case Study 12: Monte Carlo-Based Traffic Simulation in

Baton Rouge, Louisiana .........................................................26612.4.1 Data Preparation and Program Overview ................26612.4.2 Module 1: Interzonal Trip Estimation ......................26912.4.3 Module 2: Monte Carlo Simulation of Trip

Origins and Destinations .......................................... 272

K19056_C000toc.indd 12 9/26/2014 10:44:23 PM

xiiiContents

12.4.4 Module 3: Monte Carlo Simulation of Trip Distribution ............................................................... 273

12.4.5 Module 4: Trip Assignment and Model Validation .. 27312.5 Summary ............................................................................... 276

References ............................................................................................................. 279

Index ...................................................................................................................... 293

K19056_C000toc.indd 13 9/26/2014 10:44:23 PM

K19056_C000toc.indd 14 9/26/2014 10:44:23 PM

xv

List of FiguresFIGURE 1.1 Dialog window for attribute query. .................................................. 12

FIGURE 1.2 Dialog window for spatial query...................................................... 13

FIGURE 1.3 Dialog windows for projecting a spatial dataset. ............................. 14

FIGURE 1.4 Dialog window for calculating a field. ............................................. 15

FIGURE 1.5 Dialog window for defining mapping symbols. .............................. 16

FIGURE 1.6 Population density pattern in Baton Rouge in 2010......................... 17

FIGURE 1.7 Dialog window for multiple ring buffer. .......................................... 18

FIGURE 1.8 Dialog window for the Dissolve tool. .............................................. 19

FIGURE 1.9 Dialog window for creating a graph in ArcGIS. .............................20

FIGURE 1.10 Dialog window for spatial join. ..................................................... 21

FIGURE 1.11 Population density patterns based on data at the census tract and block levels. .............................................................................22

FIGURE 1.12 Flow chart for Case Study 1: (a) Part 1. (b) Part 2. ........................22

FIGURE A1.1 Rook versus queen contiguity. .......................................................24

FIGURE A1.2 Workflow for defining queen contiguity. .......................................24

FIGURE 2.1 An example for the label-setting algorithm. ....................................30

FIGURE 2.2 Dialog window for geocoding hospitals based on geographic coordinates. ..................................................................................... 35

FIGURE 2.3 Dialog window for geocoding hospitals based on street addresses. ...................................................................................... 36

FIGURE 2.4 Dialog window for an attribute join. ............................................... 37

FIGURE A2.1 A valued-graph example. .............................................................. 41

FIGURE A2.2 Dialog window for defining Toolbox properties. .......................... 43

FIGURE A2.3 The Google Maps API tool user interface for computing O-D travel time matrix. ................................................................44

FIGURE A2.4 Estimated travel time by ArcGIS and Google. ............................. 45

FIGURE 3.1 The floating catchment area (FCA) method for spatial smoothing...... 48

FIGURE 3.2 Kernel density estimation. ............................................................... 49

K19056_C000e.indd 15 9/26/2014 11:01:25 PM

xvi List of Figures

FIGURE 3.3 Zhuang and non-Zhuang place names in Guangxi, China. .............54

FIGURE 3.4 Dialog window for summarization. ................................................. 55

FIGURE 3.5 Zhuang place name ratios in Guangxi by the FCA method. ........... 57

FIGURE 3.6 Kernel density of Zhuang place names in Guangxi. ....................... 57

FIGURE 3.7 Spatial interpolation of Zhuang place names in Guangxi by the IDW method. ............................................................................. 58

FIGURE 3.8 Population change rate in Baton Rouge 2000–2010. ....................... 63

FIGURE 3.9 Flow chart for implementing the TDW method. ............................. 63

FIGURE 4.1 Constructing Thiessen polygons for five points. .............................. 71

FIGURE 4.2 Reilly’s law of retail gravitation. ................................................. 72

FIGURE 4.3 Proximal areas for the Cubs and White Sox. ..................................77

FIGURE 4.4 Probability of choosing the Cubs by the Huff model (β = 2.0). ...... 81

FIGURE 4.5 Proximal areas for public hospitals in Louisiana. ...........................84

FIGURE 4.6 Service areas for public hospitals in Louisiana by Huff model (β = 3.0)............................................................................................86

FIGURE 4.7 Probability of visiting LSUHSC-Shreveport Hospital by Huff model (β = 3.0). ................................................................................87

FIGURE A4.1 Interface for implementing the Huff model. ................................. 91

FIGURE 5.1 The basic floating catchment area (FCA) method in Euclidean distance. ...........................................................................................96

FIGURE 5.2 The two-step floating catchment area (2SFCA) method in travel time. .......................................................................................98

FIGURE 5.3 Conceptualizing distance decay in G2SFCA: (a) Gravity function, (b) Gaussian function, (c) 2SFCA, (d) E2SFCA, (e) kernel, and (f) three-zone hybrid approach. .................................. 101

FIGURE 5.4 Flow chart for implementing the 2SFCA in ArcGIS. .................... 104

FIGURE 5.5 Accessibility to primary care physician in Chicago region by 2SFCA (20-mile catchment). ......................................................... 105

FIGURE 5.6 Accessibility to primary care physician in Chicago region by 2SFCA (30-min catchment). .......................................................... 106

FIGURE 5.7 Comparison of accessibility scores by the 2SFCA and gravity-based methods. .............................................................................. 109

FIGURE A5.1 Interface for implementing G2SFCA method. ............................ 112

FIGURE 6.1 Regional growth patterns by the density function

approach. ....................................................................................... 119

K19056_C000e.indd 16 9/26/2014 11:01:25 PM

xviiList of Figures

FIGURE 6.2 Excel dialog window for regression. ............................................. 122

FIGURE 6.3 Excel dialog window for Format Trendline. ................................. 123

FIGURE 6.4 Illustration of polycentric assumptions. ........................................ 127

FIGURE 6.5 Population density surface and job centers in Chicago. ................ 130

FIGURE 6.6 Density versus distance exponential trend line (census tracts). ..... 133

FIGURE 6.7 Density versus distance exponential trend line (survey townships). ..................................................................................... 136

FIGURE 7.1 Scree plot and variance explained in principal components analysis (PCA). ............................................................................... 147

FIGURE 7.2 Major steps in principal components analysis (PCA) and factor analysis (FA). ....................................................................... 148

FIGURE 7.3 Dendrogram for a cluster analysis example. .................................. 150

FIGURE 7.4 Conceptual model for urban mosaic. ............................................. 153

FIGURE 7.5 Districts and subdistricts in Beijing. .............................................. 154

FIGURE 7.6 Spatial patterns of factor scores in Beijing. ................................... 158

FIGURE 7.7 Social areas in Beijing. .................................................................. 159

FIGURE 8.1 Mean centers and ellipses for racial–ethnic groups in the Chicago area. ................................................................................. 167

FIGURE 8.2 SaTScan dialog windows for point-based spatial cluster analysis. ..................................................................................... 171

FIGURE 8.3 A spatial cluster of Zhuang place names in Guangxi, China. ....... 172

FIGURE 8.4 ArcGIS dialog window for computing Getis–Ord General G. ..... 173

FIGURE 8.5 Clusters of homicide rates based on local Moran’s Ii. ................... 183

FIGURE 8.6 Clusters of homicide rates based on Gi∗. ........................................ 184

FIGURE 8.7 GeoDa dialog window for defining spatial weights. ..................... 186

FIGURE 8.8 GeoDa dialog window for regression. ........................................... 187

FIGURE 8.9 Standard residuals in the GWR model. ......................................... 188

FIGURE 8.10 Spatial variations of coefficients from the GWR model: (a) factor 1, (b) factor 2, (c) factor 3, (d) JA (job accessibility). ........ 189

FIGURE 9.1 Female breast cancer death rates in Illinois for 2003–2007. ......... 194

FIGURE 9.2 Example of assigning spatial order values to areas. ...................... 197

FIGURE 9.3 Example illustrating REDCAP: (a) a spatially contiguous tree is built with a hierarchal clustering method, (b) partitioning the tree by removing the edge that optimizes the SSD measure. ......... 200

K19056_C000e.indd 17 9/26/2014 11:01:25 PM

xviii List of Figures

FIGURE 9.4 Interface windows in REDCAP. ....................................................204

FIGURE 9.5 Late-stage breast cancer rates in zip code areas in the Chicago region in 2000. ...............................................................................206

FIGURE 9.6 Distribution of late-stage breast cancer rates in the Chicago region in 2000: (a) 292 zip code areas, (b) 195 new areas.............207

FIGURE 9.7 Screen shot for “Dissolve” in data aggregation. ............................209

FIGURE 9.8 Late-stage breast cancer rates in newly defined areas in Chicago in 2000. ........................................................................... 210

FIGURE 9.9 Hot and cold spots of late-stage breast cancer rates in newly defined areas in the Chicago region in 2000. ................................ 211

FIGURE A9.1 User interface of the MLR method. ............................................ 214

FIGURE 10.1 Interaction between population and employment distributions in a city. ........................................................................................ 219

FIGURE 10.2 A simple city for illustration. ....................................................... 222

FIGURE 10.3 Spatial structure of a hypothetical city. .......................................224

FIGURE 10.4 Population distributions in various scenarios. .............................228

FIGURE 10.5 Service employment distributions in various scenarios. .............228

FIGURE A10.1 Interface of the Garin–Lowry model tool (with external distance table). ...........................................................................234

FIGURE 11.1 TAZs with employment and resident workers in Columbus, Ohio. ............................................................................................. 243

FIGURE 11.2 Interface of R. ..............................................................................246

FIGURE 11.3 Five selected hospitals in the p-median model. ........................... 253

FIGURE 12.1 Monte Carlo simulations of (a) resident workers, and (b) jobs. ...265

FIGURE 12.2 Traffic monitoring stations and adjacent areas in Baton Rouge..... 267

FIGURE 12.3 Workflow of the TSME. ..............................................................269

FIGURE 12.4 TSME interface for the intrazonal trip estimation module. ........ 271

FIGURE 12.5 TSME interface for the Monte Carlo simulation of O’s and D’s module. .................................................................................. 272

FIGURE 12.6 TSME interface for the Monte Carlo simulation of trips module. ........................................................................................ 274

FIGURE 12.7 TSME interface for the trip assignment and validation module. .... 275

FIGURE 12.8 Observed versus simulated traffic. .............................................. 276

K19056_C000e.indd 18 9/26/2014 11:01:25 PM

xix

List of TablesTABLE 1.1 Types of Relationships in Combining Tables .......................................6

TABLE 1.2 Types of Spatial Joins in ArcGIS .........................................................9

TABLE 1.3 Comparison of Spatial Query, Spatial Join and Map Overlay ........... 10

TABLE 2.1 Solution to the Shortest Route Problem ............................................. 31

TABLE 4.1 Fan Bases for Cubs and White Sox by Trade Area Analysis ............. 79

TABLE 4.2 Population by Hospital Trade Areas in Louisiana .............................85

TABLE 5.1 Comparison of Accessibility Measures ............................................ 107

TABLE A5.1 Items To Be Defined in the Accessibility Toolkit Interface .......... 113

TABLE 6.1 Linear Regressions for a Monocentric City...................................... 123

TABLE 6.2 Polycentric Assumptions and Corresponding Functions ................. 128

TABLE 6.3 Regressions Based on Monocentric Functions ................................ 132

TABLE 6.4 Regressions Based on Polycentric Assumptions 1 and 2 (n = 1837 Census Tracts) .................................................................. 135

TABLE 7.1 Idealized Factor Loadings in Social Area Analysis ......................... 152

TABLE 7.2 Basic Statistics for Socioeconomic Variables in Beijing (n = 107) ..... 155

TABLE 7.3 Eigenvalues from Principal Components Analysis .......................... 156

TABLE 7.4 Factor Loadings in Social Area Analysis ......................................... 157

TABLE 7.5 Characteristics of Social Areas (Clusters) ........................................ 160

TABLE 7.6 Zones and Sectors Coded by Dummy Variables .............................. 160

TABLE 7.7 Regressions for Testing Zonal versus Sector Structures (n = 107) ... 161

TABLE 8.1 Rotated Factor Patterns of Socioeconomic Variables in Chicago in 1990 .............................................................................................. 180

TABLE 8.2 OLS and Spatial Regressions of Homicide Rates in Chicago (n = 845 Census Tracts) .................................................................... 185

TABLE 9.1 Approaches to the Small Population Problem ................................... 196

TABLE 9.2 Descriptive Statistics for Female Breast Cancer by Zip Code and by Constructed New Areas in Chicago Metro Area in 2000 ....202

TABLE 9.3 Regression Results for Late-Stage Breast Cancer Risks in the Chicago Region in 2000 ...................................................................207

K19056_C000f.indd 19 9/26/2014 10:58:03 PM

xx List of Tables

TABLE A9.1 Items To Be Defined in the MLR Toolkit Interface ....................... 215

TABLE 10.1 Simulated Scenarios of Population and Service Employment Distributions ................................................................................... 227

TABLE 11.1 Location–Allocation Models ..........................................................250

TABLE 11.2 Service Areas for the Clinics ......................................................... 253

TABLE 12.1 Major Tasks and Estimated Computation Time in Traffic Simulation ....................................................................................... 271

K19056_C000f.indd 20 9/26/2014 10:58:03 PM

xxi

ForewordThis book introduces the reader in a gentle and unassuming way to the notion that the spatial structure of cities and regions is organized around ideas about the spatial geometry of cities in terms of distances, densities of occupation, and nearness or proximity usually referred to as accessibility. These are the driving forces of the way spatial structures defining our cities self-organize into recognizable forms and func-tions, and during the last 50 years, they have been catalogued and researched using formal methods and models that provide a unifying sense of the way the physical form of cities is organized. In general, cities grow from some central location, tradi-tionally the market place that is often established accidently or relates to some pre-dominant natural advantage such as a river crossing or harbor. But as the city grows around this pole or centre, it provides the essential structure of the city with its land uses and movement patterns, reinforcing the resulting configuration. Sometimes, when the centralizing forces are destroyed by those of decentralization, new hubs or centers emerge in the periphery—edge cities, thus generating landscapes which are polycentric, composed of multiple cores and clusters of different sizes which func-tion in an autonomous whole.

These are the theories and models that constitute the subject matter of this text. Fahui Wang provides an excellent introduction to these various models and the meth-ods that are used to link them to data and thence to prediction, but he does much more than this for he casts all the models that he introduces into a framework which is dominated by desktop GIS, specifically ArcGIS. Not only is the text an excellent and cogent summary of the main theories that explain the spatial structure of our cities, it is a working manual for making these theories operational by turning theo-ries into models that are then estimated or fitted to existing cities using the various software and extensions that have been developed in the field of GIS during the last 20 years. Readers are thus treated to a view of operational theory building and mod-eling in the social sciences with the focus on spatial structures and contemporary software which ultimately empowers the persistent reader who works through the book with tools and methods for turning theory into practice.

This revised and extended edition of Fahui Wang’s book, originally published as Quantitative Methods and Applications in GIS in 2006, is divided into three sec-tions. Section I deals with getting started with ArcGIS, which focuses on key func-tions involving mapping populations densities, computing distances and travel times, and interpolating and smoothing spatial surfaces from discrete points and areal data. This sets the scene for Section II which deals with basic quantitative methods and applications starting with defining trade areas or hinterlands which are key factors in business geography. Measuring accessibilities that depend on gravity and distance come next with applications to health care, and this is followed by linear analysis that is used to transform nonlinear functions of population density into forms that can be estimated for cities. Extracting structure from data at different scales using principal components and factor analysis with clustering comes next, and this section of the

K19056_C000g.indd 21 26-09-2014 17:48:25

xxii Foreword

book is concluded with a new chapter on spatial statistics that also builds on cluster-ing with respect to local and more global structures.

Section III of the book treats more advanced topics: regionalization which again relates to clustering, various linear methods of land use modeling and optimization, and finally, a new chapter on modeling of traffic based on Monte Carlo techniques. A very nice feature of the book is the wealth of examples that are included. Different places such as Louisiana State, Baton Rouge LA, Columbus OH and Chicago IL in the U.S., and Beijing and Guangxi Province of China, amongst other locations are used, while the applications of these various techniques are to density, retail trade, health care, hospital provision, social area analysis, the spatial incidence of cancer, and traffic.

In this book, Fahui Wang shows just how far GIS has progressed. Virtually everything that was developed prior to the GIS age can now be applied, further developed, and interpreted through the GIS lens, and in some respects, his examples illustrate the great range and diversity of potential applications which is the mark of a mature technology. In fact, GIS is becoming part of the routine tool kit that any analyst would use in studying data that varies across space. What is intriguing about the treatment here is that the edifice of theory of urban and regional systems that draws on locational analysis and social physics, although still advancing, is becom-ing increasingly integrated with GIS, and it is treatment of the subject area such as that developed here that shows how relevant these tools are to contemporary urban policy. In fact, it is this focus on policy that marks the book. Those reading it will find that the author weaves together explanations of the theories involved with their translation into tools and models and their estimation using straightforward statistics with notions about different ways of applying these models to real problems that have strong policy implications. At the end of the day, it is not only understanding cities better that is the quest for the tools introduced here but also understanding them in deep enough ways so that effective policies can be advanced and tested that will pro-vide more sustainable and resilient cities—one of the challenges of near and medium term futures in the socioeconomic domain. This book sets a standard, shows how this can be achieved, and charts the way forward.

Michael BattyCentre for Advanced Spatial Analysis

University College, London, United Kingdom

K19056_C000g.indd 22 26-09-2014 17:48:25

xxiii

PrefaceOne of the most important advancements in recent social science research (includ-ing applied social sciences and public policy) has been the application of quantita-tive or computational methods in studying the complex human or social systems. Research centers in computational social sciences have flourished in major uni-versity campuses including Harvard University (http:/ / www. iq. harvard. edu/ ), Stanford University (https://css-center.stanford.edu/), UCLA (http://ccss.ucla.edu/), University of Washington (http://julius.csscr.washington.edu/), and George Mason University (http:/ / www. css. gmu. edu/ ). Many conferences have also been organized around this theme (http://computationalsocialscience.org/). Geographic Information System (GIS) has played an important role in this movement because of its capability of integrating and analyzing various data sets, in particular spatial data. The Center for Spatially Integrated Social Science (CSISS) at UC Santa Barbara, funded by the National Science Foundation (1999–2007), has been an important force in pro-moting the usage of GIS technologies in social sciences. The Centre for Advanced Spatial Analysis (CASA) at the University College, London, UK, is also known for its leading efforts in applied GIS and geo-simulation research with a focus on cities. The growth of GIS has made it increasingly known as geographic information sci-ence (GISc), which covers broader issues such as spatial data quality and uncertainty, design and development of spatial data structure, social, and legal issues related to GIS, and many others.

Many of today’s students in geography and other social science-related fields (e.g., sociology, anthropology, business management, city and regional planning, public administration) share the same excitement in GIS. But their interest in GIS may fade away quickly if the GIS usage is limited to managing spatial data and mapping. In the meantime, a significant number of students complain that courses on statistics, quantitative methods, and spatial analysis are too dry and feel irrelevant to their interests. Over years of teaching GIS, spatial analysis, and quantitative methods, I have learned the benefits of blending them together and practicing them in case studies using real-world data. Students can sharpen up their GIS skills by applying some GIS techniques to detecting hot spots of crime, or gain better understanding of a classic urban land use theory by examining their spatial patterns in a GIS environ-ment. When students realize that they can use some of the computational methods and GIS techniques to solve real-world problems in their own field, they become bet-ter motivated in class. In other words, technical skills in GIS or quantitative methods are learned in the context of addressing subject issues. Both are important for today’s competitive job market.

This book is the result of my efforts of integrating GIS and quantitative (com-putational) methods, demonstrated in various policy-relevant socioeconomic appli-cations. The applications are chosen with three objectives in mind. The first is to demonstrate the diversity of issues where GIS can be used to enhance the studies related to socioeconomic issues and public policy. Applications spread from typical

K19056_C000h.indd 23 10/8/2014 5:10:56 PM

xxiv Preface

themes in urban and regional analysis (e.g., trade area analysis, regional growth patterns, urban land use and transportation) to issues related to crime and health analyses. The second is to illustrate various computational methods. Some methods become easy to use in automated GIS tools, and others rely on GIS to enhance the visualization of results. The third objective is to cover common tasks (e.g., distance and travel time estimation, spatial smoothing and interpolation, accessibility mea-sures) and major issues (e.g., modifiable areal unit problem, rate estimate of rare events in small population, spatial autocorrelation) that are encountered in spatial analysis.

One important feature of this book is that each chapter is task-driven. Methods can be better learned in the context of solving real-world problems. While each method is illustrated in a special case of application, it can be used to analyze dif-ferent issues. Each chapter has one subject theme and introduces the method (or a group of related methods) most relevant to the theme. For example, spatial regres-sion is used to examine the relationship between job access and homicide patterns; systems of linear equations are analyzed to predict urban land use patterns; linear programming is introduced to solve the problem of wasteful commuting and allo-cate healthcare facilities; and Monte Carlo technique is illustrated in simulating urban traffic.

Another important feature of this book is the emphasis on implementation of methods. All GIS-related tasks are illustrated in the ArcGIS platform, and most sta-tistical analyses are conducted by SAS. In other words, one may only need access to ArcGIS and SAS in order to replicate the work discussed in the book and conduct similar research. ArcGIS and SAS are chosen because they are the leading software for GIS and statistical analysis, respectively. Some specific tasks such as spatial clus-tering and spatial regression use free software that can be downloaded from the Internet. Most data used in the case studies are publicly accessible. Instructors and advanced readers may use the data sources and techniques discussed in the book to design their class projects or craft their own research projects. A link to the website is provided for downloading all data and computer programs used in the case studies (see “List of Major GIS Datasets and Program Files”). I plan to run a blog under my homepage (http://ga.lsu.edu/faculty/fahui-wang/) in support of the book.

The book has 12 chapters. Section I includes the first three chapters, covering some generic issues such as an overview of data management in GIS and basic spa-tial analysis tools (Chapter 1), distance and travel time measurement (Chapter 2), and spatial smoothing and interpolation (Chapter 3). Section II includes Chapters 4 through 8, covering some basic quantitative methods that require little or no pro-gramming skills: trade area analysis (Chapter 4), accessibility measures (Chapter 5), function fittings (Chapter 6), factor analysis (Chapter 7), and spatial statistics (Chapter 8). Section III includes Chapters 9 through 12, covering more advanced topics: regionalization (Chapter 9), a system of linear equations (Chapter 10), linear programming (Chapter 11), and Monte Carlo simulation (Chapter 12). Sections I and II may serve an upper-level undergraduate course. Section III may be used for a graduate course. It is assumed that readers have some basic GIS and statistical knowledge equivalent to one introductory GIS course and one elementary statistical course.

K19056_C000h.indd 24 10/8/2014 5:10:56 PM

xxvPreface

Each chapter focuses on one computational method except for the first chapter. In general, a chapter (1) begins with an introduction to the method, (2) discusses a theme to which the method is applied, and (3) uses a case study to implement the method using GIS. Some important issues, if not directly relevant to the main theme of a chapter, are illustrated in appendixes. Many important tasks are repeated in dif-ferent projects to reinforce the learning experience (see “Quick Reference for Spatial Analysis Tasks”).

My interest in quantitative methods has been very much influenced by my doc-toral advisor, Jean-Michel Guldmann, in the Department of City and Regional Planning of the Ohio State University. I learned linear programming and solving a system of linear equations in his courses on static and dynamic programming. I also benefited a great deal from the mentorship of Donald Haurin in the Department of Economics of the Ohio State University. The topics on urban and regional density patterns and wasteful commuting can be traced back to his inspiring urban econom-ics course. Philip Viton, also in the Department of City and Regional Planning of the Ohio State University, taught me much of the econometrics. I only wish I could have been a better student then.

It has been 8 years since the publication of the previous edition of this book. Several friends and many users, especially my students, have found some errors and given me valuable feedbacks. Several tasks in the previous edition had to be imple-mented in the deprecated ArcInfo workstation environment (and its associated AML program). The software, particularly ArcGIS, has advanced with increasing capaci-ties and more user-friendly interfaces. Numerous people have urged me to automate some popular tasks in Python. I have also collected a few new tricks from my own research that I hope to share with a broader audience. The reasons for preparing a new version just kept piling on. It came to a point that I could not find any more excuse for not doing it when my sabbatical request was approved by the Louisiana State University in early 2013.

Inevitably I have lost a step or two over the years. The added administrative duty did not help either. I truly felt that the revision took more work than the previous edi-tion. I used most of my summer and the sabbatical in the fall of 2013 on the book, and often found myself working on the book in my office very early in the mornings of weekdays and weekends for much of the spring semester of 2014. In this version, several popular tasks such as trade area delineation, spatial accessibility measures (2SFCA), mixed-level regionalization (MLR), and Garin–Lowry model are auto-mated as a convenient toolkit in ArcGIS; and some common tasks such as REDCAP regionalization, wasteful commuting measure, and Monte Carlo simulation are now implemented in more user-friendly programs. All case studies have been tested mul-tiple times, and instructions were based on ArcGIS 10.2. I would like to believe that it was the amount of revision work that merited the investment of so much of time. It could be an understatement to say that more than 60% of the materials (including case studies) are new. I hope that the readers will be convinced likewise that it is a worthy cause.

I am so grateful for the generous help from several individuals in preparing this second edition. Yujie Hu coded the program TSME used in the case study in Chapter 12 and coauthored Chapter 12. He also tested all case studies and corrected numerous

K19056_C000h.indd 25 10/8/2014 5:10:56 PM

xxvi Preface

errors. Haojie Zhu implemented the Python programs for three toolkits (trade area delineation, spatial accessibility measures, and Garin–Lowry model). Both Yujie and Haojie are PhD students in the Department of Geography and Anthropology, Louisiana State University. Lan Mu in the Department of Geography, University of Georgia, implemented the mixed-level regionalization method in Python and coau-thored Appendix 9B in Chapter 9. Xinyue Ye in the Department of Geography, Kent State University, coauthored Appendix 10D on cellular automata for urban land use modeling in Chapter 10. Carmi J. Neiger of Elmhurst College tested the case stud-ies in Chapters 1 through 4 and edited the four chapters. I thank Michael Batty for graciously writing the new Foreword on short notice.

Finally, I would like to thank the editorial team at Taylor & Francis: acquisition editor, Irma Britton; production editor, Laurie Schlags; and many others, including typesetters, proofreaders, cartographers, and computer specialists. Thank you all for guiding me through the whole process.

This book intends to mainly serve students in geography, urban and regional planning, public policy, and related fields. It can be used in courses such as (1) spa-tial analysis, (2) location analysis, (3) quantitative methods in geography, and (4) applications of GIS in business and social science. The book can also be useful for social scientists in general with research interests related to spatial issues. Some in urban economics may find the studies on urban structures and wasteful commut-ing relevant, and others in business may think the chapters on trade area analysis and accessibility measures useful. The case study on crime patterns may interest criminologists, and the one on regionalization and cancer analysis may find the audience among epidemiologists. Some basic GIS knowledge (e.g., one introductory GIS course) will help readers navigate through the case studies smoothly, but is not required.

Additional material including the datasets and program files is available from the CRC website: http:/ / www. crcpress. com/ product/ isbn/ 9781466584723. (The final URL will be provided by electronic products department.)

K19056_C000h.indd 26 10/8/2014 5:10:56 PM

xxvii

AuthorFahui Wang is James J. Parsons Professor and Chair of the Department of Geography and Anthropology, Louisiana State University. He obtained the BS degree in geography from Peking University, China, the MA degree in economics and the PhD degree in city and regional planning from the Ohio State University. His research interests include GIS applications in human geography (urban, economic, and transportation), city and regional planning, and public policy (crime and health). His work has been supported by the National Science Foundation, U.S. Department of Energy, U.S. Department of Health and Human Services (Agency for Healthcare Research & Quality and National Cancer Institute), U.S. Department of

Housing and Urban Development, U.S. Department of Justice (National Institute of Justice and Office of Juvenile Justice and Delinquency Prevention), and National Natural Science Foundation of China. He is on the editorial boards of several inter-national journals including Annals of the Association of American Geographers, Annals of GIS, Applied Geography, and Chinese Geographical Science. He has pub-lished over 100 refereed articles.

K19056_C000i.indd 27 9/25/2014 6:29:24 AM

K19056_C000i.indd 28 9/25/2014 6:29:24 AM

xxix

List of Major GIS Datasets and Program Files*

* http://www.crcpress.com/product/isbn/9781466584723

Case study Data folder GIS dataset Program file Study area

1 BatonRouge BR.gdb, Census.gdb

East Baton Rouge Parish

2 Louisiana LA_State.gdb, LAState.gdb

Louisiana

3A China_GX GX.gdb Guangxi, China

3B BatonRouge BR.gdb East Baton Rouge Parish

4A Chicago ChiRegion.gdb Huff Model.tbx Chicago Region

4B Louisiana LA_State.gdb, LAState.gdb

Huff Model.tbx Louisiana

5 Chicago ChiRegion.gdb Accessibility.tbx Chicago Region

6 Chicago ChiUrArea.gdb monocent.sas, polycent.sas, mono_twnshp.sas

Chicago Urban Area

7 Beijing BJSA.gdb PCA_FA_CA.sas, BJreg.sas Beijing, China

8A Chicago ChiUrArea.gdb Chicago Urban Area

8B China_GX GX.gdb Guangxi, China

8C Chicago ChiCity.gdb PCA_FA.sas Chicago City

9 Chicago/Chizip

Chizip shapefiles poisson_zip.sas, Mixed-Level Regionalizaiton Tools.tbx

Chicago Region

10 SimuCity SimuCity.gdb Garin-Lowry Model.tbx Hypothetical city

11A Columbus Columbus.gdb WasteComm.R, LP.sas Columbus, Ohio

11B BatonRouge BR.gdb East Baton Rouge Parish

12 BatonRouge BRMSA.gdb TSME.exe, Features To Text File.tbx

Baton Rouge MSA

K19056_C000j.indd 29 09-10-2014 14:42:11

K19056_C000j.indd 30 09-10-2014 14:42:11

xxxi

List of Quick Reference for Spatial Analysis Tasks

TaskSection First

Introduced (Step)Figure

IllustrationSections (Steps)

Repeated

2SFCA method (automated toolkit) Appendix 5B Figure A5.1

Areal weighting Interpolation Section 3.5.1 (1–3)

Attribute join Section 2.4.1 (5) Figure 2.4 Section 3.3.1 (3), Section 4.3.2 (6)

Attribute query Section 1.3.1 (1) Figure 1.1 Section 3.3.1 (4)

Dissolve Section 1.3.2 (10) Figure 1.8

Euclidean distance matrix computation

Section 2.4.1 (3) Section 3.3.1 (2), Section 4.3.2 (5)

Geocoding by address locator Section 2.4.1 (1) Figure 2.3

Geocoding by geographic coordinates Section 2.4.1 (1) Figure 2.2

Geographically weighted regression (GWR)

Section 8.7.2 (8)

Getis–Ord general G Section 8.7.1 (3)

Huff model (automated toolkit) Appendix 4B Figure A4.1

IDW spatial interpolation Section 3.3.2 (10) Section 6.5.1 (2)

Kernel density estimation (KDE) Section 3.3.2 (9) Section 6.5.1 (2)

Local Gi∗ Section 8.7.1 (4)

Local Moran’s Ii (LISA) Section 8.7.1 (4)

Mean center Section 8.2 (1)

Moran’s I Section 8.7.1 (3)

Network dataset building Section 2.4.2 (8) Section 10.3 (1, 3)

OLS regression in ArcGIS Section 8.7.2 (5) Section 6.5.1 (4-5)

Projection of spatial data Section 1.3.1 (3) Figure 1.3

Proximal area by Euclidean distance Section 4.3.1 (2–3)

Proximal area by travel time Section 4.4.1 (1–3)

Spatial join Section 1.3.2 (13) Figure 1.10

Spatial query Section 1.3.1 (2) Figure 1.2 Section 6.5.1 (1)

Spatial regression in GeoDa Section 8.7.2 (7)

Standard deviational ellipse Section 8.2 (3)

Standard distance Section 8.2 (2)

Summarize attribute Section 3.3.1 (5) Figure 3.4 Section 4.3.1 (4), Section 4.3.2 (7)

Travel time O-D matrix computation Section 2.4.2 (9–12)

Section 10.3 (1, 3), Section 11.2.2 (3)

Weighted centroids Section 4.3.1 (1)

K19056_C000k.indd 31 09-10-2014 15:05:10