Project Meeting 2021.02.04

5b/6a Task 1 Project Management (milestone)

AMPO update
- Welcome Caitlin Cook, AMPO technical programs director
- AMPO starting back up the modeling working group
Contracting topics
- Any final feedback on the partner agreement or funding agreement send in this week
- RSG needs to return the bench contract
Academic cooperation
- @Joe to reach out to Xuesong about his DTALite integration email
- @Joe to reach out to Greg Macfarlane about his UTRAC research grant email
Some potential PRs possibly coming soon from the SANDAG cross border model project:
- make scheduling probability lookup model generic for trips and tours
- ability to run trip mode choice within tour mode choice to get trip model choice logsums for inbound and outbound trips assuming a chosen tour mode
- ActivitySim wants to support SANDAG's use of the platform for the cross border model
- New functionality would be optional and needs to include a test example
- We'll talk more about it on a technical call with Wu
- Would be good to share materials ahead of time for review since its not a straightforward topic

Bill and Shimon reviewed bay area data to inform model design
OBS has fare medium and fare category but not subsidy
CHTS has fare subsidy information
TNC survey has questions related to commute benefits (provided by your employer)
Clipper card registered users database could maybe tell us about percent of card holders that are seniors, kids, employer provided, etc.
Clipper card data might not be very representative
Clipper start program with a 50% discount for low-income is a new program
Some research by Greg Erhardt showed some populations absent from the Clipper card data
Our Consortium budget for this task is limited so we want to be mindful of the ask of Joel/SEMCOG
@Shimon/Bill to summarize findings and follow-up with Joel
DaySim's ability to assert/override fare subsidy should be included in the model design for policy testing and what-if type analysis
The disconnect between what data we have and what we want from the model may be an issue

I reviewed the PRs and merged the multiprocessing revision to support populationmsim
We discussed the trip number of participants PR and added a few good ideas to the discussion thread. The key is for the platform to be flexible and support needs - to have configurable options