This year our team is looking to use a scouting application based purely on data from thebluealliance.
First, I’m wondering how quick that data is available on TBA after a match. At the end of the day? An hour after the match? After the entire competition? I’m guessing this varies from regional to regional so if it helps, we’ll be using this at St. Louis regional.
Second, in order for the application to be useful, I’ll have to perfect it for the 2019 season. I was wondering if anyone knew when the documentation for Match_Score_Breakdown_2019 will be available.
Slightly untrue. TBA polls FIRST for data. Polling happens every minute, but FIRST has ~2 minute cache on their API. So in the worst case, TBA data is delayed from when FMS commits scores by up to 3 minutes, or 1.5 minutes on average.
Edit: Just because I was curious, here is a distribution of how long it took TBA to get updated data from FIRST after a match finished in 2018. We don’t track this directly*, so I compared the time a match was last updated in our database to the time a match started + 2 minutes and 30 seconds. Because adding a video to a match also updates a match in our database, I limited the results to only show matches that were last updated within 15 minutes of finishing. This removes outliers due to videos as well as event internet going down, etc., but it also means that the histogram shows only a subset of matches from 2018.
* Actually we do have a way of tracking this directly, but that required way more work.
FIRST publishes the official schema for the data (which TBA uses), so it’ll be available Soon™.
Typically though, it’s been made available closer to the end of build season in February.
Any reason why this typically takes until February for FIRST to release the API info? I would suspect teams would be more likely to integrate it into their scouting programs if it was available on kickoff. I’d like to give my students the experience of connecting to an API, but releasing that info right around bag day makes it tough as we are typically doing final testing of the scouting program at that point.
I think it may be tied to annual FMS updates? I’m not entirely sure when those begin their development cycle (I’m guessing it’s after the game gets finalized), but I’m guessing the code that generates the data needs to be rewritten for each game.
Of course, that doesn’t explain why they can’t just give us the schema at kickoff, I can’t see it being that up in the air come that point.
Perhaps somebody who knows more about FMS/these things can provide insight?
