Exchanging predicted occupancy in GTFS

[skinkie]

In GTFS-RT a transit agency can exchange occupancy information. I would like to pull the https://developers.google.com/transit/gtfs-realtime/reference#enum-occupancystatus into GTFS, specifically as column in stop_times.txt, to describe the predicted occupancy between the current stop to the next stop.

[devadvance]

Given the lengthy, pre-existing debate about the value of those specific enums (see #223), do you have more information on the type of occupancy you would wish to convey?

[skinkie]

I would like to send a prediction with that enum beforehand. The discussion of what values the enum contains is irrevant. At this moment we have in use: no information (the null value), empty, many seats available, few seats available, standing room only, full.

[derhuerst]

I would like to send a prediction with that enum beforehand. The discussion of what values the enum contains is irrevant.

But if GTFS-RT adopted the mentioned OccupancyStatus enum and #223 then came to the conclusion to remove/change it, we would have 2 (instead of 1) changes breaking existing implementations. Am I misunderstanding something here?

[skinkie]

In my opinion it should be the same enum. Regardless of the values that that enum at that moment in time contains. But if we are talking about a complete removal (where it changes to a percentage) this obviously should become the occupancy attribute that matches the occupancy attribute in GTFS-RT.

[skinkie]

Someone at mobility data is working on it, would be great if the discussion actually took place first.
#240

[harshverma2000]

Given this discussion, I would like to know if there is any thought on the lead time for such predicted occupancy information being related to Riders...

Rail Agencies would be more interested in knowing when a Predictive Value or Prediction is available for Riders in addition to seeing just the Real-time value.
Would it be of value to show this 5 minutes earlier, 3 minutes earlier or 7 minutes earlier? How would Latency affect this?

The important thing is to support riders to be able to know where to stand to board the train when standing at the platform and which car to board is finally going to drive their need.

The internal model for the Predictive Value may not need to be the decisive factor…

[skinkie]

The case for prediction is in the exploration planning phase days before travel. Hence for this scheduled exchange changes during the day are not suggested. If any update should take place, this should happen in GTFS-RT.

[harshverma2000]

The case for prediction is in the exploration planning phase days before travel. Hence for this scheduled exchange changes during the day are not suggested. If any update should take place, this should happen in GTFS-RT.

Thanks for your comments! We are also thinking this should happen in GTFS-RT. Where would you suggest we should take up this discussion in GTFS-RT?
For the Static, we agree the value should be an average of Months+ or more and averages of days of the week.