The pipeline makes it easy to create synthetic populations and simulations for other regions than Île-de-France. In any case, we recommend to first follow instructions to set up a synthetic population for Île-de-France and (if desired) the respective simulation. The following describes the steps and additional data sets necessary to create a population and simulation for Lyon and surroundings.
Lyon is not included in the census data set that is uesd for Île-de-France
(Zone A). Instead, Zone E needs to be obtained from the same source. Download the dbase version of Zone E and put the
respective file (FD_INDCVIZE_2015.dbf) into the data/rp_2015 folder.
Only if you plan to run a simulation (and not just generate a synthetic population),
you need to obtain additional data from OpenStreetMap.
Geofabrik provides a cut-out for the former Rhône-Alpes region. Download the region file in .osm.pbf format and put the file into the
folder data/osm.
Again, only if you want to run simulations, the digital transit schedule is required. Unfortunately, there is no consolidated GTFS schedule avaiable for the region of interest. Hence, it is necessary to collect all relevant GTFS schedules one by one. Here, we provide a selection of links, which is not necessarily exhaustive (for instance, it would be possible to add schedules for Saint Etienne or Grenoble)
Download all the zip'd GTFS schedules and put them into the folder data/gtfs.
Afterwards, you should have the following additional files in your directory structure:
data/rp_2015/FD_INDCVIZE_2015.dbf
Only for simulation:
osm/rhone-alpes-latest.osm.pbfgtfs/GTFS_TCL.ZIPgtfs/export_gtfs_voyages.zipgtfs/export-intercites-gtfs-last.zipgtfs/export-ter-gtfs-last.zip
Note that the file names may change slightly over time as GTFS schedule are updated continuously.
To generate the synthetic population, the config.yml needs to be updated. While
the relevant code points to the Île-de-France data sets by default, you can
adjust the paths inidividually. To let the pipeline use the Zone E census
data set, add the following to config.yml in the config section:
config:
# ...
census_path: rp_2015/FD_INDCVIZE_2015.dbf
# ...Furthermore, by default the pipeline will filter all other data sets for the Île-de-France region. To make it use the selected region, adjust the configuration as follows:
config:
# ...
regions: []
departments: ["01", 38, 42, 69, 69M] # 26 "07"
# ...This will make the pipeline filter all data sets for the departments noted in the list above, which is a set of the closest departments around Lyon. If you want to generate the whole (ancient) Rhône-Alpes region, add the commented out department identifiers to the list.
You can now enter your Anaconda environment and call the pipeline with the
synthesis.output stage activated. This will generate a synthetic population
for Lyon and surroundings.
To prepare the pipeline for a simulation of Lyon, the paths to the OSM data sets and to the GTFS schedule must be adjusted explicitly:
config:
# ...
gtfs_path: gtfs/GTFS_TCL.ZIP;gtfs/export_gtfs_voyages.zip;gtfs/export-intercites-gtfs-last.zip;gtfs/export-ter-gtfs-last.zip
osm_path: osm/rhone-alpes-latest.osm.pbf
# ...Note that the pipeline will automatically cut GTFS and OpenStreetMap data to the relevant area (defined by the filter above) if you run the simulation.
For a region as large as the one in this example, we recommend to futhermore reduce the complexity of the road network. You can define the smallest possible road level like this:
config:
# ...
osm_road_level: tertiary
# ...The pipeline will filter out automatically roads with a lower classification.
To test the simulation and generate the relevant MATSim files, run the pipeline
with the matsim.output stage enabled.