GTFS

[amandaha8]

• Transit Bunching: figuring out how to fix timestamps after converting to a datetime object. There are 23 hours between two time stamps of the same date and only 1.5 hours apart when I manually inspect a route that was ran by the City of Visalia. Figuring out ways to throw away extreme data.
• Agency Level Metrics: calculating metrics on this grain in the middle of the pipeline instead of at the end. I have now reached the end of the testing stage with 2024-only dates and next week I hope to polish this off.
• GTFS Digest: Trying to prevent the kernel from dying after a left merge is performed segment speed portion of gtfs_digest/merge_data.py. The geodataframes merge/concat just fine, but it seems like the memory dies when trying to export out the dataset into GCS.

[github-actions]

nbviewer URLs for impacted notebooks:

• gtfs_digest/31_transit_bunching_v2.ipynb
• gtfs_digest/34_segment_speeds.ipynb
• rt_scheduled_v_ran/11_agency_agg.ipynb