Project Meeting 2021.12.09

Agenda

• Code and model implementation review of vehicle type model

  • Created vehicle_type_choice.py model code and vehicle_type_choice configs in mtc
  • Alternatives are created programatically based on the "combinatorial_alts" variabel in the model yaml file
  • Model spec has been updated to include both choosers and alternatives data which simplifies the expressions
  • Estimation data bundle generation is already included in the code (probabilities for model 4 are not included)
  • Might want to restructure / generalize the on-the-fly creation of combinatorial alternatives code at some point in the future for use in other models
  • Model option 2 is currently being included in the test

• Check in on project requirements for the visualizer

  • Added concept of "slicers" to create discrete bins for continuous variables
  • Added concept of "aggregate" to combine alternatives for comparison (e.g. all transit modes aggregated into one "transit" output for the plot)
  • Added preprocessor config file to calculate things like skim distances
  • Request for setting "automated" break points, e.g. by quantile, 10 evenly spaced bins, etc.
  • For proper brushing (selection across multiple plots), the summary csvs would have to be much more combinatorial and might change the server requirements. Implementing this functionality is outside the current scope, but there is a lot of interest

• RFP review and questions

  • Deadline for questions in writing is tomorrow, Friday Dec 10

• Check in on GitHub issues

  • Suggest everyone to review their created issues and see if they are still relevant
  • Feel free to change/add tags