data science student coursework help

[The_Watcher]

Hi there,
Strange request here, and i suppose this is the right place for it:

1) I'm a part-time MSc Data Science student. One of my modules is all about data programming with Python;
2) I've been set a very broadly worded piece of coursework, deadline end-June;
3) I'd quite like to try my hand at building from scratch a straightforward Elo rating system for 1v1 rated FAF games. I'm by no means a Python expert so I think this will stretch me about as far as my current competencies go;
4) Just wondered if there might be some way I could get my hands on a bunch of records showing something along the lines of:
[Player A]
[Player B]
[Winning Player]
[Map]

Many thanks if anyone could point me in the right direction. I can see the FAF client maintains a 1v1 Rating leaderboard, but it's not obvious to me any of the underlying match records are publicly available.

I've not actually played this game in several years but I'm an avid Gyle cast watcher... hence "The_Watcher"

Cheers all!

[FtXCommando]

We have an open API that allows you to query our data except for things we mark as personal information.

https://github.com/FAForever/faf-java-api

[Axle]

There's an old csv dump of game results here: https://github.com/Axle1975/pytrueskill ... playdumper

(and a c++ implementation of trueskill with python bindings in the rest of the repo)

[The_Watcher]

thanks guys that's really helpful

[Katharsas]

THe API is using JSONAPI to allo you to dynamically request the exact fields from the tables you need.

Query games:
https://api.faforever.com/swagger-ui.ht ... e/get_game

endTime, name, replayTicks, replayUrl, startTime, validity, victoryCondition, featuredMod, host, mapVersion, playerStats, reviews, reviewsSummary

Query game results:
https://api.faforever.com/swagger-ui.ht ... layerStats

afterDeviation, afterMean, ai, beforeDeviation, beforeMean, color, faction, result, score, scoreTime, startSpot, team, game, player

However note that you should probably restrict your dataset to a somehwat small timeframe, because the data in those fields may not be consistent over big timeframes.