SIRI-ET feed quality overview

[leonardehrenfried]

After the NeTEx feed for 2025 is available again, I tried again with the following modifications to the data:

• Run an XSLT sheet on the netex data so that the ids of the ScheduledStopPoints match the ones in SIRI (SIRI and NeTEx stop point IDs do not match #215)
• ScheduledStopPoint support in OTP (not yet merged)

The result is this a success rate of 18%. That is not a good rate.

Here is the relevant log message:

363 of 1991 update messages were applied successfully (success rate: 18.23%)
988 failures of errorType=NOT_MONITORED
369 failures of errorType=EMPTY_STOP_POINT_REF
257 failures of errorType=NO_FUZZY_TRIP_MATCH
2 failures of errorType=NEGATIVE_HOP_TIME
12 failures of errorType=NEGATIVE_DWELL_TIME

We have the following obvious errors:

• NOT_MONITORED: the update doesn't contain realtime information and is discarded (SIRI feed contains large number of journeys that don't have any realtime information #214)
• EMTPY_STOP_POINT_REF: the update contains an empty stop point ref (SIRI feed contains empty stop point refs #213)

They affect over half of the updates.

Then we have errors of NEGATIVE_HOP_TIME and NEGATIVE_DWELL_TIME. This means that the SIRI data contains time traveling, which OTP of course rejects.

The rest is data where the fuzzy trip matcher could not identify the correct trip. We may be able to improve that but it will always remain a guessing game.

cc @rcavaliere

[leonardehrenfried]

The good news is that I was able to match 363 of 1991 updates.

[leonardehrenfried]

Today I tried to find some result which contained some realtime information and managed to do so:

But from a superficial check, coverage is not great.

[rcavaliere]

@leonardehrenfried did you use https://otp.opendatahub.testingmachine.eu? Can you provide me some more detailed info about your attempts, so that I can also have a look?

[leonardehrenfried]

Yes, it's all deployed to now.

For me I get the most real-time data when I use local busses in Merano, Bressanone or Bolzano.

Here are some examples:

Merano

https://tinyurl.com/28gdv4f5

Bressanone

https://tinyurl.com/2betgmfh

I never get real-time data for trains.