The purpose of this repository is to perform the intermediate modeling steps in the preparation of a persona-based large-scale agent-based transport simulation for the Île-de-France region in France.
The synthetic population pipeline can be generated as described in the branch personas of
https://github.com/eqasim-org/ile-de-france
To use this repository, two files are needed:
input/RP2019_INDCVIZA_csv.zipshould be the 2019 census data for region A (Île-de-France)data/clusters_2019.csvshould be a file indicating the baseline persona of every observation in the 2019 census (one column calledID)
The notebook 01 Scenario calibration.ipynb will calibrate the distributions of household size, number orf household cars and location type (wokring in same municipality as home, working in same department, ...) to match certain predefined targets that are given on top of the notebook.
The output are four files, one per scenario, that describe the distributions. The files can be used as input to the population synthesis process.
Based on the calibrated distributions, a new distribution of personas across the population should be obtained that "generates" these distributions. The process is performed in 02 Persona calibration.ipynb.
The output are four files, one per scenario, that describe the persona distribution. The files can be used as input to the population synthesis process.
After running the pipeline, one can use 03 Validation.ipynb to validate that all the marginals have been taken into account properly in the scenario generation. Note that population growth is included in those generated scenarios.
The scenarios should be put into output/scenarios, so output/scenarios/idf_1pct_s1_persons.csv, for instance. The scenarios should be ["baseline", "projection", "s1", "s2", "s3", "s4"].