Coming soon: non-revenue service data and uncertainty values

[Developer at MBTA]

Dear Developers,

In order to streamline our internal data pipeline and enable improvements to rider-facing information, we will begin publishing non-revenue service in our data streams on Thursday, December 7.

This message a preliminary alert to help facilitate a smooth transition. Early adjustments to your implementations (if needed) are recommended to accommodate this update — and if you have any questions, we encourage you to reach out ahead of the change.

Note: At this time, non-revenue trips on light rail already appear in the feeds as revenue trips. We are currently working to identify and flag non-revenue service on light rail.

Here is what non-revenue service will look like in our data feeds:

GTFS-Realtime

Non-revenue trips will only be published to our GTFS-Realtime enhanced JSON feeds. You can view our GTFS documentation for this change here.

Non-revenue trips:

1. Will not have a schedule-based, revenue GTFS trip_ID.
   1. The trip_ID will follow a NONREV-XXXX format.
2. Include all stops with a ScheduleRelationship of skipped. 

v3 API

In order to reduce confusion, we are implementing an explicit filter within the API: revenue_status.

1. The revenue_status filter is available under the /trips, /predictions, and /vehicles resources.
2. The filter is set to revenue by default.
   1. Setting the filter to revenue only outputs revenue service.
   2. Setting the filter to non-revenue only outputs known non-revenue service.
   3. Setting the filter to all outputs all service.

You can view our API documentation PR for this change here.

Uncertainty values:

We will also surface uncertainty values, when available, in our data streams, as specified in the StopTimeEvent message documentation. These metrics assess the reliability of real-time bus and subway predictions and will be accessible as follows:

• For Bus and Subway Predictions within GTFS-RT:
  • Our GTFS-RT feeds already include prediction uncertainty values.
  • Detailed documentation on these values, per mode, is available here.
• Within the v3 API:
  • uncertainty values will be surfaced in the v3 API starting December 7. They will be included in arrival and departure predictions as arrival_uncertainty and departure_uncertainty, respectively.

As always, if you have any questions, please don’t hesitate to reach out.

Thanks,

Developer @ MBTA