Research Request - Can we leverage gtfs-segments in our pipeline?

[tiffanychu90]

Complete the below when receiving a research request, and continue to add to this issue as you receive additional details and produce deliverables. Be sure to also add the appropriate project-level label to this issue (eg gtfs-rt, DLA).

Research Question

Single sentence description: Bus Performance Metrics webinar by WMATA surfaced some interesting repos...and gtfs-segments seems like something we could use (presentation showed route-direction segments between 2 stops, with stop_id1, and stop_id2 that are basically equivalent to our stop_pair column when we aggregate to route-direction.

• Epic - Daskify RT segment speeds #592
• switch back to vp-sjoined-to-roads, aggregate shapes up to route-direction with stop_pairs #965 (when stop_pair column is first introduced)

Detailed description:
Quick poke through repos shows that gtfs-segments is based on gtfs_functions. It also looks like gtfs-segments doesn't use classes and underlying functions take a somewhat-processed df or gdf. Should be straightforward to use these functions.

• https://github.com/jtl-transit/rove - bus performance dashboard ingesting GTFS, AVL, APC, and other ops data
  https://mobility.mit.edu/sites/default/files/TRR__Transit_Performance_and_Journey_Visualization_Dashboard.pdf - academic paper about the ROVE dashboard
• https://github.com/valhalla/valhalla - Routing engine package, multiple presenters used it to clean up GTFS route shapes and align to OSM
  https://journals.sagepub.com/doi/10.1177/03611981211009226 - 'Using GTFS to Calculate Travel Time Savings Potential of Bus Preferential Treatments'
• https://pypi.org/project/gtfs-segments/ - GTFS segments package
• https://github.com/Bondify/gtfs_functions - GTFS functions package
• https://github.com/sophspilio/toolingtransit - NoVa Transportation Commission Transit Resource Center projects. For the R users (but also contains some python notebooks)
• Slack thread

How will this research be used?

• Let's see what we can leverage from these packages, if it's faster, more accurate, and more well-maintained than what we can do in our scripts.

Deliverables

• Exploratory notebook to make sure segments are similar to the way we use stop segments
• Scripts in rt_segment_speeds/scripts/ and consolidate cut_normal_segments.py and cut_special_segments.py into 1 script that uses gtfs-segments