Project Meeting 2025.04.08 - ActivitySim/activitysim GitHub Wiki

Agenda

  • Telecommuting design options

Technical

  • Comparisons of SANDAG production and SANDAG example model
    • There are minor differences between
    • There is no significant differences that suggest taking telecommuting in a different direction
  • Telecommuting design options
    • Presentation can be found here: https://drive.google.com/drive/folders/14v7yFqGrczmrFR7ltrsVWgRgEQiBUnCh
    • Current ActivitySim models
      • Telecommute frequency model, which is then used as covariate in downstream models
    • Work completed to date
      • Empirical analysis of survey data
      • Software enhancements: added Telecommute status model
    • Design considerations
      • Represent "hybrid" work patterns
      • Represent multiple work activities and locations
      • Describe in-home work activity schedules
      • Model sequencing
    • Options for representing "hybrid" work patterns
      • Expand telecommute status model
        • Replace telecommute status with Telecommute Arrangement
        • Alternatives: In-home work activities only; in-home + out-of-home activities; Out-of-home activities onl
      • Add "also out-of-home work activity" model
      • Expand coordinated daily activity pattern model
        • Add w/ in-home vs w/in-home patterns to existing DAP options
      • Use the existing mandatory tour frequency model - considered not an option based on previous discussions
      • Hybrid
        • Pro: Allow telecommuters to engage in an out-of-home work tour
          • According to PSRC, make up significant percentage of telecommuters
        • Con: Would need to select some arbitrary observed duration for an in-home work activity to "count" for this discrete choice; Need to re-estimate mandatory tour scheduling for "hybrid" workers
    • Represent multiple work activities and locations
      • Expand telecommute status model
        • Pros: Generalization of, and therefore superior to, "hybrid" suggestion; Explicitly represents multiple work activities and locations
        • Cons: Need to set a threshold for duration of "in-home" work activity; No data on more than one in-home activity; Re-estimate mandatory tour scheduling for those with >= 1 in-home work activity
    • Scheduling At-home Work Activities Options
      • Add in-home work activity duration model
      • Add in-home work activity scheduling (i.e., start time and duration) model
      • Add joint strategic work day arrangement model
        • Don't have data to pursue this option
      • Use the existing mandatory tour scheduling model
      • Pros: Software to explicitly represent at-home work activities; Sets the stage for more consistent and realistic itineraries
      • Cons: No data on in-home start and end times; Need to decide which to schedule first, in-home or out-of-home; Need to re-estimate mandatory tour scheduling
    • Recommended Solution: Multiple Work + Scheduling
      • Telecommute arrangement model: Addresses hybrid and multiple work activity
      • In-home Work Activity Scheduling
        • Software gives users the following options: 1. Duration Only. No blocking of time windows for other activities; 2. Start time and duration. Block resulting window from other activities
      • Duration outcome can inform downstream models
      • CDAP options: whether Telecommute arrangement model happens before, after, or in-conjunction with CDAP model
        • Recommend: CDAP then Telecommute Arrangement
    • Next Steps
      • April 17 - Submit draft recommendations memo
      • April 24 - Follow-up discussion on design, leading to a decision
    • Discussion
      • New Gallop Pole: Remote capable jobs and employees; Would be represented in the telecommute frequency model and the work from home model
        • SANDAG uses 2-digit NAICS code as variables in the model
      • None of the CDAP options are that ideal
  • Thursday's meeting is cancelled