rGTFS

Tool to compare planned GTFS with real GTFS (rGTFS) extracted from GPS data.

Working API

import rgtfs

# Generage Realized Trips from GTFS
rgtfs.generate_realized_trips_from_gtfs(gtfs_path)

# Treat Rio de Janeiro BRT realized trips
rgtfs.helpers.treat_rj_brt_realized_trips(brt_raw_realized_trips_path)

New Tables Documentation

Realized Trips

Field Name	Type	Required	Description
vehicle_id	ID referencing vehicles.vehicle_id	Required	identifies vehicle
departure_datetime	Datetime	Required	Time at which the vehicle departs from the first stop
arrival_datetime	Datetime	Required	Time at which the vehicle departs from the stop
departure_id	ID referencing stops.stop_id or garages.garage.id	Required	Departure unique identifier. Can be a stop_id or a garage_id.
arrival_id	ID referencing stops.stop_id or garages.garage.id	Required	Arrival unique identifier. Can be a stop_id or a garage_id. Can be empty if trajectory_type is garage_to_stop.
distance	Float	Required	Distance travelled in the trajectory in seconds
elapsed_time	Integer	Required	Trajectory duration in seconds
average_speed	Float	Required	Trajectory average speed in km/h
trajectory_type	Enum	Required	One of the followin trajectory types: 1. complete_trip: A complete one-way trip. Departure stop and Arrival stop should map to the first and last stop of the trip respectively. 2. not_complete_trip: An incomplete one-way trip. Departure stop should map to the first stop of the trip, but the trip was sundelly aborted so it has no Arrival. 3. garage_to_stop: A trajectory between a garage and a stop or otherwise. One of the stops should map to a garage_id.
trip_id	ID referencing trips.trip_id	Optional	Trip unique identier. Only applicable for trajectory_type: complete_trip and not_complete_trip
trip_short_name	Trip name	Optioanl	Public facing text used to identify the trip to riders, for instance, to identify train numbers for commuter rail trips. If riders do not commonly rely on trip names, leave this field empty. A trip_short_name value, if provided, should uniquely identify a trip within a service day; it should not be used for destination names or limited/express designations. It can be used if trip_id is not available.

How to go about it

Set standards of how the GPS data should look like. What are the accepted columns? How should the feed be structured? What is the best fit with GTFS?

What are the expected results?

• Retrospective GTFS, rGTFS, which is a GTFS of what actually happend on date YYYY-MM-DD

  • Fully dependend on GPS. It'd have to create the shapes, stops, lines, etc
  • Based on planned GTFS. It'd use the current GTFS shapes, stops and lines to support the algorithm

• Comparisson between GTFS and rGTFS

  • Difference in bus frequency/waiting times
  • Difference in the fleet. What are the lines larges fleet difference?
  • Ponctuality score.

How the module should look like?

• io (read and write to GTFS)
• GTFS object
• rGTFS builder to GTFS object

Similar Projects

Related projects by other people:

• https://github.com/CUTR-at-USF/retro-gtfs/tree/GTFS-Realtime (a fork of this code extending it to GTFS-realtime data sources)
• https://github.com/WorldBank-Transport/Transitime