Rocket League Analytics

Rocket League detailed stat dashboard with expected goals analysis.

See dashboard here: https://sertalpbilal.github.io/rocket_league_analytics

Generating xG model and dashboard

The xG model file is under src/xg_model. The model depends on having a valid key from ballchasing.com API.

Steps:

Setting up environment
Add your BallChasing token to .env.sample file and rename it to .env
Collecting random game data (optional)
Navigate to src/scripts and run python collect_model_games.py
This step will download latest 1000 ranked double games under data/model directory.
Generating xG model (optional)
Run python generate_xg.py to populate xG model.
Note that you might need to run this inside a Docker container if you cannot install carball package.
This step will populate two files under data/model: xg.model (gradient boosting xG model) and xg.scaler (feature scaler)
Download your games
You can go to scripts folder and run python download_my_games.py
This step will download your games to data/replay and will populate data/json and data/dataframe directories for future use.
Generate the dashboard
Finally you can go to src and run python boxcartest.py to populate the dashboard.

Screenshots:

Considering all games

Click here to view in an alternative format

Considering the latest streak of games

Click here to view in an alternative format

Output:

Alongside all of the charts produced by the program, the program also outputs 11 tables.

Tables for the latest streak of matches can be found here.