Calculate travel time between zone pairs (travel time matrix)

[Hussein-Mahfouz]

• [x] 1. What times of day do we want to calculate travel time matrices for?
  • I think at least 3 different times (morning peak, afternoon off-peak, and evening peak). Comparing accessibility levels at different times of day is useful to examine temporal component of DRT service provision
• [x] 2. What modes do we want to calculate travel time for?
  • public transport
  • car
• [x] 3. What gtfs feeds do we need?
  • Bus: I have downloaded gtfs data for all of england from the Bus Open Data Service
  • Rail*: Where do we get the data from? It will need to be converted using UK2GTFS - see #7
• [x] 4. calendar.txt to harmonise feeds. (e.g. if bus and rail feeds are for different time periods) - https://github.com/Hussein-Mahfouz/drt-potential/issues/9
  • one approach to give all feeds the same start and end date - example code
  • how do we sync the feeds but maintain the service dates of the different agencies?

(*) The importance of rail depends on the geographic region (e.g. Leeds or Leeds City Region (LEP) or West Yorskhire)

[Hussein-Mahfouz]

Functions done here f5386cc09a9d9bea1842fa2e84969ae7ffc809ab.

If I want travel distance, then I will need to use the detailed itineraries function. This function doesn't accept gtfs feeds with frequencies.txt files, so I will have to add a preprocessing step to: (a) convert frequency to stop_times (b) remove frequencies.txt

The bods gtfs has both frequencies and stop_times files. It is important that the frequencies_to_stop_times() function does not overwrite any existing stop_times.txt - it should add to it

[Hussein-Mahfouz]

I use gtfstools::frequencies_to_stop_times() here. I've decided not to run the detailed_itineraries function for now . It may be necessary if I adopt the methodology outlined in Option 2 here