Project Meeting 2021.02.18 - ActivitySim/activitysim GitHub Wiki

Task 1 Project Management (milestone)

  • No contracting updates
  • Let's work on diversifying our contributor based, see issue #379
    • Both agency staff and consultant staff
  • I updated the wiki a bit and added multi-region-timing google sheet
    • Need to add machine specs to google sheet
  • Doyle merged all the latest branches to develop2 and I need to review and then we'll PR to develop
  • Debrief of meeting with Greg at BYU about small urban area model
    • Oregon DOT interested in a small MPO simple Asim setup as well
    • @Joe to follow-up with Ohio DOT since maybe interested as well
    • Can Oregon DOT or others earmark additional funds to AMPO just for this purpose?
    • @Joe to check with AMPO but a separate sidebar agreement seems possible and others can join if they want
  • Debrief of meeting with Xuesong about DTALite improvements
    • Long term there's interest in adding network supply models to the effort
    • DTALite has been used successfully by several DaySim users for maz maz all streets shortest paths
    • How does this model handle tolls? We think it might
    • Lots of good stuff coming out of Xuesong's shop so let's keep working together
  • RSG setting up an ActivitySim model for CMAP
    • Will migrate the CT-RAMP model to ActivitySim
    • CMAP wants a simpler / better supported model
    • CMAP encouraged to join the consortium
    • I'll connect CMAP with the PMC and they are invited to join the meeting if they want
    • As the community grows, so does the management effort so let's encourage agency membership
  • @Joe to email AMPO about setting up an AWS machine setup for testing our examples with a consistent accessible solution

Task 2 Complete Estimation Mode (milestone)

  • Newman working on notebooks with support from Doyle for improvements as needed

Task 3 Telecommute Model in Cooperation with SEMCOG (milestone)

  • Plan to circle back with consortium once SEMCOG project has results to share
  • Work from home looks fine, but telecommute freq looks funny because occupations not available
  • SEMCOG won't have person occupation but has industry so planning to revise/re-estimate
  • Should the official ActivitySim example be the SEMCOG example or something else (i.e. with occupations)? We're thinking it should be the SEMCOG example since its probably too much to maintain two examples (and especially one with no clear owner/input data)
  • This relates to our plan to support/maintain multiple examples
  • Need to get working on the task 6 memo to help facilitate the conversation

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

  • Joel/team to review data

Task 5 Performance Enhancements (milestone)

  • Doyle done with taz/maz location sampling
  • Stefan to now test and to review results for goodness-of-fit
  • Representative logsums next

Task 6 Prototype multiple models test system (milestone)

  • We're getting the examples in order for the improved test system
  • I need to draft the proposed plan and one topic to discus is the level of detail required when validating/verifying model results with PRs
  • The document will discuss possible tiers of completeness
  • And we can build an inventory of issues/topics/ideas to help inform everyone's understanding of the needs and possible solutions

Task 7 Input checking and error handling and pipeline improvements (milestone)

  • Nothing to report

Task 8 Maintenance and Support (milestone)

  • I'll email Clint to request he present on the ARC model implementation on an upcoming Tuesday call