Web conference notes, 2022.10.27 (MDS Working Group)
MDS Working Group
- Every other Thursday at 9am PT, 12pm ET, 5/6pm CET
Zoom Registration Link: https://us02web.zoom.us/meeting/register/tZAscOmhpjIuHNakPx6CNbACpjUjw1Gsucr4
One tap mobile: +19294362866,,84170989462#,,,,*612987# US (New York) - though we encourage Zoom
Note: Attendees register upon entry into the Zoom meeting. An attendee count will be posted here after the meeting:
25 Attendees
Main Topics
Unification Changes
- MDS Agency and Provider Unification - Marie Maxham, Lacuna
WGSC Meeting Organizers
- Host: Sebastian Berthaud, Blue Systems
- Facilitator: Marie Maxham, Lacuna
- Outreach: Michael Schnuerle, OMF
- Note taker: Michael Schnuerle, OMF
- trip_id - Conditionally Required instead of Required, add null note in Telemetry Data
- ask everyone about beta Geography Driven Events - contact Ride Report
- Ok to change GeoJSON points to lat/lon points?
- Provider feedback on the events current/historic, and shape of telemetry
Pain point for providers to have 2 similar but different APIs
Pulled common items up from Provider and Agency
Moved to general_information for draft
Some overlap didn't align completely
Trips
Should it stay as GeoJSON
Telemetry is GPS points instead of GeoJSON in draft. Is this OK?
Events
Greatest alignment so far
Could remove provider_name if it's redundant
Use Telemetry instead of Point? Change to lat/lon instead of GeoJSON or vice versa?
Do we want event_geographies? From Austin and Ride Report.
Telemetry Data
Should it be GPS points or GeoJSON without extra GPS data?
Filtering based on accuracy is important
Could move to points, easier to ingest into a database, away from GeoJSON, but want to know if anyone is using it for some reason.
Michael OMF and Alex SFMTA and Matt Populus agree. Parsing for DB adds a step
Vehicles
Should this be Devices?
Trips Query
New query parameter for 'route' to not return route data for privacy or data payload size.
Events (called Status Changes now)
Need to have both historic or recent/realtime events
Proposing /events/recent and /events/historical
Status changes made to address historic data processing time
Pierre Blue Systems - bug on data flow, two week threshold is a bit of a pain when 3 weeks hits and you have to look into 2 endpoints to make sense.
Brooks SFMTA - question of fixing records if issue. How to handle resending same records with an update? Maybe a Trip ID would help? Need to know what needs to be removed or added on DB.
Marie - two week barrier is an arbitrary and a bit limiting. Would like to combine into one endpoint, but need feedback. Provider feedback needed here.
Matt Populus - uses batch concept for data management, which helps. We switched to hourly downloads and disallowed pagination in Provider API - anecdotally the API was more stable as they use 1.1 and 1.2.
Do we need to keep recent events? Utility who uses it? Maybe using it as a real time concept.
If you poll vehicles you can miss events. But if no one is using Recent events, then could eliminate it.
Michael - seems a bit risky to
Telemetry Query
Specify start/end time and trip ID to get that data only.
Distinction between recent and historic telemetry? Events can inform that.
Stops basically the same across both
Reports
Didn't try to address pushing reports into Agency yet.
Batch question from SFMTA asking Populus. Downloads multiple days at once with a batch id, then if update delete old batch by ID and add new batch.
Nividya SANDAG - How far back do providers keep data? Usually in terms of service. EG LA uses 24 hours delay. Would like to document how to specify early and late time.
CHAT
Steve Brining (Blue Systems) to Everyone (12:03 PM)
Me to Everyone (12:06 PM)
Nice to meet you Pierre
Hi Tony from Detroit
Angela Giacchetti to Everyone (12:07 PM)
Good to see you both virtually!
Marie Maxham (Lacuna) to Everyone (12:07 PM)
Welcome!
Alex Demisch (SFMTA) to Everyone (12:34 PM)
I agree with Michael on point data --- it would be easier if we just had the lat/long data in fields rather than geojson. Parsing it just adds a step
Matt Davis (Populus) to Everyone (12:37 PM)
Think it's probably the same for us at Populus. It's not hard for us to munge a lat/lon (or series of them) into a geospatial object and the simplified representation is nicer to work with.
Nivedya Madankara Kottayi - SANDAG to Everyone (12:43 PM)
Also, any thoughts on how we specify the timeline for the data to be saved/shared by the provider?
Brooks Jessup (SFMTA) to Everyone (12:50 PM)
@matt I'm interested in your batch concept. Can you say a bit more about how you use it in your record correction process?
Alex Demisch (SFMTA) to Everyone (12:58 PM)
Thanks for your work and this great overview Marie!
