Project Meeting 2021.01.28 - ActivitySim/activitysim GitHub Wiki

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

  • Please get any additional MOA feedback in by next week's call
  • Trying to keep contract language generic and avoid agency specific tweaks if possible
  • AMPO has a new technical programs manager, Caitlin

5b Task 2 Strategic Planning (milestone)

  • I summarized updates to V3 of the plan
  • Incorporated everyone's feedback and edits
  • @Joe to review and see if a new partner, Rebekah, would like to review as well in order to get a new perspective

6a Task 2 Complete Estimation Mode (milestone)

  • Lots of progress discussed on Tuesday
  • Should the larch code live in the asim repo or larch repo?
  • We decided on the asim repo, with the idea that we'll be able to support multiple estimation tool integrations one day, with larch being the first
  • We're open to having ALOGIT estimators in the repo for example as well
  • It's also difficult for testing code revisions across repos and we want the estimation functionality to have test coverage
  • @Newman to build out the remaining notebooks
  • PMC ok with shifting hours between RSG and CS as long as there are no cost or scope revisions

6a Task 3 Telecommute Model in Cooperation with SEMCOG (milestone)

  • SEMCOG working on setting it up

6a Task 4 Transit Pass Ownership Model in Cooperation with SEMCOG (milestone)

  • Discussed rough draft design / thinking on Tuesday but still work to do before agreeing to a design
  • Certainly don't want a simple 50% discount assumption like some other models
  • @Joel going to look into the data and produce some summaries to help answer questions
  • Let's try to do data summaries across multiple regions as well
  • Can we look into the bay area data? @Bill to investigate
  • Caltrain, bart, muni all have different pass programs, and your employer is a big part of your likelihood of having a pass

6a Task 5 Performance Enhancements (milestone)

  • Discussed some profiling results on Tuesday
  • We think the PSRC trip purpose and destination issue is likely data related, @RSG to help PSRC resolve under support task
  • The taz/maz sampling speedup seems like a no-brainer
  • But we need more info to understand the representative logsum idea, including LOE
  • What about other ideas? Strings to factors, etc? Let's make sure to consider those too
  • @I'll talk with @Jeff about drafting a list of improvements and LOEs for discussion next week
  • Can we speed up the accessibility model some more as well? DaySim does some precalculation. We can and should, but its a significant task that will need some design/discussion since there's a number of different accessibility calculators out there and we'll want to think about asim's plans for the future

6a Task 8 Maintenance and Support (milestone)

  • Need to incorporate PRs and then release again (with updated estimation functionality and accessibility improvements too)