multi day GTFS flattened to 1 day supply leading to overestimating supply

[AntoineGrapperon]

Hi,
The tutorial on Raleigh is actually using a multiday GTFS and assigning the 1 day demand on the multiday supply that is the GTFS. (Liza just confirmed me that)

3 suggestions to fix this:

• in the tutorial notebook, possibility to add some GTFS data processing to extract only 1 day (then the burden of processing the GTFS relies on the user)
• include a parameter in fast-trip Run method which allows the user to select the day. Then, the data processing to extract the day specific data is included in fast-trip code. This allows you to drop most of the GTFS and focus on the day o f interest.
• consider the timestamp (date + time) instead of just the time. This option is probably the most accurate (because it allows you to have trip that start at 11.30pm and finish at 1 am (ok, long trip, but you see what I mean). For big areas, it is also consuming a lot of memory...

The calendar.txt table tells you if your service is active, you also want to look at calendar_dates.txt to make sure the day of study is not some kind of special day.

Hope that helps,
Antoine