Started 2016-06-28 and ended 2018-09
The idea is to use archived real-time vehicle location data to retroactively create a perfectly accurate GTFS dataset: the GTFS of an omniscient god – but because we are human, achievable only in hindsight.
I wrote the software to accomplish this largely in Python and based the initial code on the NextBus API. All data comes directly from the API and after a while observing the movements of transit vehicles, the program can spit out a retrospective GTFS archive in the standard CSV format.
(The project has since been forked by some lovely folks at CUTR to allow it to work with GTFS-realtime data sources as well.
This work comes, to some degree, out of my master’s thesis which used similar methods and very similar data sources, though it extends those techniques quite a bit. Retro-GTFS has now been used as the basis of a number of academic papers and was presented at the 2016 GIScience conference in Montreal.
Component of: Accessibility Beyond the Schedule
Causal links to the following projects:
Causal links from the following projects: