Data to gather

[EwoutH]

A short list of data to gather to answers research questions.

Global

• Mode choice distribution
• Traffic volume
• Average speed (Analyzer: Return average_speed and vehicle_density with area_to_pandas toruseo/UXsim#121)
  • Needs Collect data only from part of network toruseo/UXsim#122 to be really accurate
• Average delay

Per hour

• Mode choice distribution (2678d86)

Per hour, per area

• Traffic volume (622ab49, Analyzer: Add time-filtering functionality toruseo/UXsim#123)
• Vehicles parked (813e598)
• Average speed (622ab49, Analyzer: Add time-filtering functionality toruseo/UXsim#123)
  • Does have a bug, for some reason this is negative.
• Average delay (622ab49, Analyzer: Add time-filtering functionality toruseo/UXsim#123)

Per Agent

• Economic gain (time/money saved due to faster travel)

toruseo/UXsim#120 would help a lot with per-hour data gathering.

[EwoutH]

Mode choice distribution per area is a bit difficult, because you have OD-pair. You could do origin or destination, but for now just collect it by hour.

[EwoutH]

Was an especially though nut to crack, but we do now have Traffic volume and Average delay per hour, per area.

Vehicles parked should be easy (or at least easier) now that we have (3d7f1a7). That's the last big one, and then we can run a base case and start experiments.

Those together should deliver some beautiful spatial choropleth animations.

CC @quaquel

[EwoutH]

Parked cars are tracked continuously, simple implementation as expected: 813e598. Can be sampled as often as required.

One thing is that our OD data is the aggregate of a day, and doesn't make a distinction between rush hours in different directions. So that does limit the usefulness of this a little.

It could still emerge somewhat from population data though. If at one district very much people live, and at another many work, they will make their first trip to that direction. But it won't be the full effect.