Project Meeting 2021.04.13

Technical Meeting

• Issues review
• Discuss updated transit pass ownership model
  • Joel reviewed the bay area data
  • Clipper card is not really a pass so can we remove it?
  • How does each type of pass relate to the fare discount per trip?
  • Average subsidized fare is about 50% but this is very region specific
  • Revised model alternatives - subsidized pass, not subsidized pass, no pass
  • In application, the pass would be segmented by markets and each can have a separate fare discount amount
  • Can we try to figure out what they actually paid in the data for each trip? / figure out the discount
  • Maybe the outcome isn't discrete....can we make it continuous?
  • There's a lot of variety in this topic across regions and within and the data is incomplete
  • The HH survey is not enough data granularity
  • ARC OBS has many many pass types
  • What is the out of pocket cost for each actual trip? It's difficult to say
  • How to calculate the discount, that's the key
  • SEMCOG doesn't have the data to reestimate this model
  • Next steps, reestimate without Clipper card
  • Iterate on design with team, post design into a Google doc
  • Hold on implementation for now
  • Model will influence auto ownership and mode choice
• Release just about ready to go, still need to document pre-sampling feature and settings file inheritance, #411
• Task 7 usability issues plan:
  • Improve RAM/chunking settings, #406
    • Switch from chunk_size parameter to available_memory parameter (say 56GB) and let ActivitySim figure out the appropriate chunk_size given the number of processors specified
  • Print info about no alternative found, #404
  • Turn off restarting if desired, #405
• Discuss ActivitySim related presentations at the TRB Planning Applications and/or AMPO Annual conference?