olliestanley / autoscout Goto Github PK

fbref has added carries & progressive carries

Add to util.load_csv as an alternative to Pandas DataFrames

Package installable from PyPI in longer term once a more complete set of features is available

Adding GitHub Actions will allow automated linting and unit testing runs

Formatting tools (black, isort, flake8, mypy) can all use configurations to define a consistent style

Use kloppy for provider-agnostic event data analysis

Add examples notebooks in examples/ directory with outputs

Also add .gitattributes to ignore Jupyter Notebooks as a repository language

Support a special case of comparison radars where the comparison record is the dataset average

Although many functions are data dependent it is still possible to unit test significant portions of the code base

In autoscout/data/fbref.py, support downloading match-level data for both teams and players. This must be achieved in a way which does not induce request spam, to ensure fbref terms of use are not violated.

For team and player aggregate-level data, implement functions in search.py to find teams and players by statistics

Opta data via WhoScored cannot be scraped as it is against the Terms of Use

This issue will be updated and worked on if a usable event data source is identified

rating_inputs.json and all radar/*.json files are affected

fbref includes wage data for players. Support downloading this in autoscout/data/fbref.py.

Relevant resources:

• https://fbref.com/en/comps/9/wages/Premier-League-Wages

For team and player match-level data, implement a preprocess function to possession adjust statistics.

Blockers: #1