A project to learn a model of bus arrival time using time series data published by SEPTA.
Right now, I am just working on retrieving the route shape data for a bus route, and using using it to normalize the lat/lon position of a bus into route-centered coordinates (ie. between 0 and 1 from the start to the end of the route).
Install pygtfs. Change line 166 in pygtfs/gtfs_entities.py to be
_validate_wheelchair = _validate_int_choice([None,0,1,2], 'wheelchair_boarding') to allow for null wheelchair_boarding fields. Then use the command-line gtfs2db tool to make a sqlite database with the contents of the Septa GTFS bus feed.
Install lewfish/gis-tools and make sure it's on the PYTHONPATH.
Then in this project, invoke pip install -r requirments.txt
In lewfish/septaprediction/septa.py change the value of septa_fn to point to the database file created with gtfs2db.
You can run a test by invoking python -m lewfish.septaprediction.septa.