public-transport-network

Analysis of public transportation systems using network science.

Raw data

The analysis is based on the public infrastructure data of St. Petersburg, Russia. The two sources of raw data are:

• data/raw/spb_routes.csv - data on public transport routes (bus, trolley, tram). Downloaded from this page.
• data/raw/osm.zip* - data from OpenStreetMap. Downloaded via Overpass API using the following queries:
  1. St. Petersburg (relation 337422):

  [out:json];
  area(3600337422)->.a;
  (
    node(area.a);
    way(area.a);
    relation(area.a);
  );
  out;
  

  2. Leningrad oblast (relation 176095):

  [out:json];
  area(3600176095)->.a;
  (
    node(area.a);
    way(area.a);
    relation(area.a);
  );
  out;
  

  * The file size is about 250 MB, therefore it cannot fit inside a Git repo. You can download it using this Google Drive link.

Further pipeline

All procedures are executed in the corresponding jupyter-notebooks (directory pipeline/) in the order suggested by their corresponding names (i.e. pipeline/1_data-preprocessing.ipynb).

Dependencies

Note that this repo (unfortunately) uses both pip and conda for package management, since installing cartopy without conda is a pain in the ass, and I also like to use my own little package myutils, which cannot be installed using conda directly from Github. The simplest way to use the repo is to run

pip install --r requirements.txt

and then additionally

conda install cartopy