Post on 23-Jul-2020
The MATSim Santiago open data model
Development and first applications
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Leonardo Camus Alejandro Tirachini Benjamin Kickhöfer Credits to: Daniel Hosse Kai Turner Felix Zwick
DLR.de • Slide 1
1. History of MATSim Santiago
2. Recent developments
3. Challenges and first applications
4. Summary and outlook
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Agenda
DLR.de • Slide 2
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
History of MATSim Santiago
DLR.de • Slide 3
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
How it all started
DLR.de • Slide 4
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
First impressions
DLR.de • Slide 5
1. Street network from OSM
2. Public transit supply data as GTFS
3. Travel diaries (and other stuff) from EOD2012 • Exported from Microsoft Access into *.csv • Includes some freight traffic (and other information)
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Open data sources
DLR.de • Slide 6
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
v1: first simulations
DLR.de • Slide 7
Raw data: • 60’054 individuals (~1%) • 113’591 trips
MATSim: • 42’459 individuals • 103’055 trips
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
v1: integrating public transit
DLR.de • Slide 8
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
v1: behavioral parameters and calibration
DLR.de • Slide 9
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Recent developments
DLR.de • Slide 10
• Remove agents that were interviewed in summer vacation time and on weekends
• Use expansion factors to create a 1% and a 10% sample
• Add time-dependent tolls to tollways
• Randomize activity end times using smartcard data
• Randomize activity locations using land use data
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Leo (v2): overview
DLR.de • Slide 11
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Leo (v2): randomize activity end times using smartcard data
DLR.de • Slide 12
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Leo (v2): randomize activity locations using land use data
DLR.de • Slide 13
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Felix (v3): implementing colectivos
DLR.de • Slide 14
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Felix (v3): travel times, slack times, fares
DLR.de • Slide 15
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Felix (v3): colectivo (also) as PT feeder
DLR.de • Slide 16
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Challenges and first applications
DLR.de • Slide 17
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Influence of sample size on travel times
DLR.de • Slide 20
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Cordon pricing schemes: setup
DLR.de • Slide 21
τin = $6.000 τout = $3.600
τin = $6.000 τout = $2.650
Outer cordon Inner cordon
Total travel distance car
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Cordon pricing schemes: elasticities 7:30-8:30
DLR.de • Slide 22
In- and outflow
Number of car trips
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Summary and outlook
DLR.de • Slide 23
• Documentation: https://svn.vsp.tu-berlin.de/repos/public-svn/publications/vspwp/2016/16-02/ (more to come for v2, v3, …)
• Different versions of the runnable MATSim scenarios (no MATSim installation necessary): https://svn.vsp.tu-berlin.de/repos/public-svn/matsim/scenarios/countries/cl/santiago/
• Code: https://github.com/matsim-org/playgrounds/tree/master/santiago
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Open data scenario
DLR.de • Slide 24
• Ideal for BSc or MSc theses; the idea is to integrate every improvement into the current state and create a new version.
• Possible improvements: • Synthesize a 10% or 100% population • Calibrate travel times, counts • Add tolled roads • Activity distribution according to land use data • Network corrections with automatic feedback to OpenStreetMap • Add freight transport (important for emissions!) • Add colectivos, taxis • Include capacity constraints for PT vehicles • Map PT routes to the road network (interaction with cars) • Include bike as transport mode • …
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Further steps
DLR.de • Slide 25
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Backup
DLR.de • Slide 26
• Social value: better transport planning through competition (identification of weak spots of the system, policy studies, environmental/social analysis, provision of public infrastructure)
• Commercial value: creating a platform for innovative mobility-based services (car sharing systems, supply chain/location planning, delivery/logistic planning, navigation, …)
• Transparency:
• Participation/engagement:
> MATSim user meeting @hEART 2017 > The MATSim Santiago open data model: Development and first applications • Benjamin Kickhöfer • 11.09.2017
Opportunities
DLR.de • Slide 27