Welcome to City Chef, here you can build your own city.
City Chef is a collection of methodologies for essentially just randomly generating data, but where the data generating process (DGP) produces outputs, with distributions a real city. The code is intended for use in an interactive (ipython) environment.
The original motivation behind this project was to provide some synthetic data for this cool project and paper.
Check out my article for a detailed walk through the project.
City Chef can synthesis cities with the following features:
Facilities - activity locations, such as households, workplaces and hospitals
Road Networks
Road Transit Routes
Rail Transit Network and Routes
Zones - statistical zones based on the city density
Population - agents, with consistent household attributes
Activity Plans - simple activity based plans for each agent
The motivation behind this project is to fill a gap in availability of city data - because of privacy issues, commercial constraints, or because the quality and quantity of data required is not available. We also use City Chef for making small data for tests or toy examples, where we'd rather not use full city scale data. Included in this project are two such example applications:
Build a city with a road network and generate bus routes. Apply spatial noise to the networks. Output the road network to OSM format. Output the bus routes to GTFS.
We use this generator for making test case data for our big network combining tools.
2. Census & Travel Survey Generator
Build a full city with facilities, statistical areas, road and transit networks. Use this city to generate households of persons with complex underlying distributions of attributes. Output household travel survey data with simple activity and mode choice. Output census marginal statistics. Output commuter OD matrix.
We use this generator for population synthesis experiments.
python3.7 ipython7.0.1
after git clone:
cd citychef
pip3 install -r requirements.txt
pip3 install -e .
This project is WIP and more a collection of methods than a API, so we apologise in advance for environment pains.
pyproj: The GeoPandas library requires pyproj as a dependency. This can be a bit of a pain to
install. For Mac OSX, activate the environment City Chef lives in and run the following commands
before installation:
pip3 install cython
pip3 install git+https://github.com/jswhit/pyproj.git
On Windows, pre-compiled wheels of pyproj can be found on this page. Manually install the correct pyproj wheel within your environment using pip.
- no doc strings
- very little logging
- no tests
- needs epsg conversions (osm and gtfs)
- needs better structure/API
- many of the classes and methods share features
- bad module names
- speed vs detail trade-off
- accessibility calculation is simplistic
- activity plan synthesis is only simple tour based
- facility locations only contain a single facility
- transit route synthesis is pretty bad, especially for trains
More broadly City Chef uses directed tree like causation for the DGPs. This is primarily for simplicity/speed. But means that circular causation is only synthesised one way, for example, the feedback between facility locations and network.