multiple buses running the same trip ID in a single day?

93 views
Skip to first unread message

Anthony Townsend

unread,
Apr 20, 2021, 7:56:07 AMApr 20
to mtadeveloperresources
We have been noticing that more than one vehicle operates using the same GTFS trip id on a single day, on a fairly regular basis, and have had to construct a new, unique identified concatenating gtfs trip id + service date + vehicle id to distinguish them.

Anyone else notice this and what the cause is–wildcatting or operator error or ?

bell...@gmail.com

unread,
May 13, 2021, 2:23:05 PMMay 13
to mtadeveloperresources
We at Lyft have noticed this too and it's causing data quality issues because of unexpected conflicts that force us to drop some data. Specifically, we've noticed it for trip updates only - vehicle positions don't seem to have this problem.

Anecdotally, I've also hung around near the "Bethune St/Greenwich St" start of the M11 route and checked the MTA Bus GTFS-RT feed vs. reality. It seems like there are many more real-time ETAs for those early stops than actual buses that show up. This is causing our app to display false real-time ETAs for those stops. I'm wondering if the issue is related - is it possible there are extra trip updates accidentally being published for MTA buses?

MTA friends, please help and let me know if concrete examples would be helpful!

Paul

sm12...@gmail.com

unread,
May 14, 2021, 8:22:11 AMMay 14
to mtadeveloperresources
Is this with the static or real time data?

Anthony Townsend

unread,
May 14, 2021, 8:30:22 AMMay 14
to mtadeveloperresources
My original post was in reference to the SiriVehicleMonitoring endpoint of the Siri API



--
You received this message because you are subscribed to a topic in the Google Groups "mtadeveloperresources" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/mtadeveloperresources/toLN7qtp3tU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to mtadeveloperreso...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/mtadeveloperresources/d620db65-0291-4a76-8521-8cda8ebc624cn%40googlegroups.com.

Paul Bellora

unread,
May 14, 2021, 9:28:04 AMMay 14
to mtadevelop...@googlegroups.com
Interesting, in our case we’re looking at the GTFS-RT feed.

sm12...@gmail.com

unread,
May 14, 2021, 9:54:45 AMMay 14
to mtadeveloperresources
I've noticed the same problem on the bus gtfs_rt. I've solved the problem by using the vehicle id, in addition to the trip id to identify an unique trip. My interest is in analyzing bus performance, after the fact. I just discard the trip/vehicle combos that don't complete the trip. I have no workaround for the people who are providing real time bus info. N.B. I'm using only the position feed. I don't know, what the trip update feed might show.

sm12...@gmail.com

unread,
May 14, 2021, 10:03:07 AMMay 14
to mtadeveloperresources
The siri feed is supposed to be derived from the gtfs_rt data, as per a sentence in the spec. One difference is that the siri feed tries to extrapolate position data based on less frequent gtfs_rt. At least, that's what's in the manual. So, repeating the same mistake should not be surprising.

I have not used the siri feed, so I cannot verify thisn.

On Friday, May 14, 2021 at 9:28:04 AM UTC-4 bell...@gmail.com wrote:

Paul Bellora

unread,
May 14, 2021, 10:44:23 AMMay 14
to mtadevelop...@googlegroups.com
Do you mean that you’re seeing multiple vehicle positions with the same trip ID? I’m only observing trip updates that conflict that way. When you have a moment, could you provide an example?

Jeremy Shaffer

unread,
May 25, 2021, 6:20:42 PMMay 25
to mtadeveloperresources
I am working with historic data collected from the SiriVehicleMonitoring endpoint of the Siri API and am seeing that more than one vehicle using the same trip ID on a given service date is quite common... Here is an example:

Route: Bx19
Service Date: Feb 1, 2021

For all of the unique trip IDs logged on the Bx19 route with service date Feb 1, 2021:
  • ~82.23% were logged by 1 vehicle
  • ~14.46% were logged by 2 different vehicles
  • ~2.48% were logged by 3 different vehicles
  • ~0.41% were logged by 5 different vehicles
  • ~0.41% were logged by 8 different vehicles

Unknown.png


Trip ID: MTA NYCT_WF_P1-Weekday-094900_BX19_313

I picked one trip ID that was logged by multiple vehicles to examine more closely: MTA NYCT_WF_P1-Weekday-094900_BX19_313. That trip ID is logged by 3 different vehicles. You can see the temporal overlap in logging in the figure below (data was collected at approximately 1 min intervals).

Unknown-2.png

Here are the vehicle locations corresponding to those records:

Unknown-3.png


You can see that 2 vehicles (293 and 295) share the same trip ID for a substantial amount of time (~1 hr), whereas the 3rd vehicle (674) logs that trip ID only briefly.

Is this the sort of issue that others are seeing?
Reply all
Reply to author
Forward
0 new messages