It seems that most of the stop_time_update of the feed contain the field schedule_relationship set to NO_DATA, which means there is no RT information for that stop.
Therefore, there are no real time predictions for 99% of the feed. Is that normal?
The current bus GTFS-R predicts the current trip. It does push out trip updates for anything that is running now or due to run in the next 60 minutes. Most of these trips would therefore not yet have any real time data and so the timetable is presented which flags stops as NO_DATA. It was 75% NO_DATA last checked. The expectation once the realtime predictions are in place for the next trip is around the 40% NO_DATA. The work for the predictions is scheduled for September (this month).
Thanks for your answer. I don’t know if it’s exactly 99%, but its definitely way bigger than 75%.
Most of the trips don’t have realtime information. It’s even tough to find one with predictions among all the others.
What’s the point on pushing all the trips within 1h range if we’re sure there is no prediction for them? The GTFS-rt should complement the GTFS static data. By adding this constant “overload” the GTFS-rt protobuf file becomes huge.
Unfortunately that is correct. The realtime predictions are only available on the running services for now. The update later this month will provide predictions on the ‘next’ services.
The reasoning behind the ‘whole’ protobuf file is based on some historic feedback. It’s been questioned if this is still required but if there are changes it will take a while before if flows through in the feed.