A software system for visualizing and analyzing transit data. Under development.

Get the sources using (modern version of) git (including submodules):

git clone --recursive https://github.com/tru-hy/transit_analysis.git
cd transit_analysis

For debian-style distribitions, you can run an installer script that tries to set the stuff up using default values. You'll need to have sudo to run it (it'll ask the password as it goes):

./install.debian.bash

TODO. See what's done in install.debian.bash.

NOTE The data importation is far from streamlined at the moment. All this should be done with one or two simple commands.

Due to Python's challenged (to be polite) import system, currently the parent of the transit_analysis directory has to be in PYTHONPATH or somehow else findable by the import mechanism by name transit_analysis to run the import scripts. You need to also have activated the virtualenv by running source .virtualenv/bin/activate.

Currently the only supported data format for the transit "metadata" is GTFS. All GTFS data that don't directly map with live data can be loaded with:

cd utils/gtfs
PYTHONPATH=../../../ ./load_common.bash <gtfs directory>

Where <gtfs directory> is directory containing the gtfs .txt files.

The system also (currently) requires version of shapes.txt which has ids specifying the (OSM) nodes corresponding to the shape points. This can be generated using gtfs-shape-mapfit and imported with:

PYTHONPATH=../../../ ./merge.bash coordinate_shape <path to noded shapes.txt>

Departures, meaning the physical drives, which are mapped to live data require additional mangling as the live data is not necessarily as such compatible with the GTFS schema. This mangling is done by "schema adapters". Currently only (working) such adapter is for City of Tampere's live SIRI data. To load those departures, run:

PYTHONPATH=../../../ ./merge.bash transit_departure <gtfs directory> ../schema_adapters/tampere.py

Currently a City of Tampere's SIRI realtime data is the only live data format supported. To import it, run

cd utils/siri
cat <siri dump file> |PYTHONPATH=../../../ ./siri_dump_parse.py load ../schema_adapters/tampere.py

NOTE The datasets are quite large and thus this will take quite a while. Especially as the parsing isn't as efficient as it could.

The dumped data needs to be mapped to the known routes for efficient visualization. This is done with:

cd utils
PYTHONPATH=../.. ./filter_routes.py filter-stop-sequences
PYTHONPATH=../.. ./filter_routes.py filter-routes

If all went well, you can issue ./rest_server.py from the project root and with default settings will find something by pointing your browser to http://localhost:8080. Create config_local.py in the source root to override the settings.py for something more suitable.

Note that there's no built in trickery to run the server in a privileged port. One, albeit somewhat risky, way to accomplish this without running the whole shebang as superuser is to grant the privileged port binding capability to .virtualenv/bin/python by commanding sudo setcap 'cap_net_bind_service=+ep' .virtualenv/bin/python.