sportsfeatures

sports-features

A library for processing sports features over a dataframe containing sports data.

Dependencies :globe_with_meridians:

Python 3.11.6:

• openskill
• pandas
• feature-engine
• tqdm
• scikit-learn
• geopy
• numpy
• pytest-is-running
• joblib
• timeseries-features

Raison D'être :thought_balloon:

sportsfeatures aims to process features relevant to predicting aspects of sporting games.

Architecture :triangular_ruler:

sportsfeatures is a functional library, meaning that each phase of feature extraction gets put through a different function until the final output. It contains some caching when the processing is heavy (such as skill processing). The features its computes are as follows:

• Process the player and teams skill levels using OpenSkill. This is an ELO like rating system giving a probability of win and loss.
• Compute the offensive efficiency of each team/player.
• Compute the time series values of the numeric features for each team/player over the various windows provided. This includes lag, count, sum, mean, median, var, std, min, max, skew, kurt, sem, rank.
• Compute the datetime features for any datetime columns.
• Remove the lookahead features.

Installation :inbox_tray:

This is a python package hosted on pypi, so to install simply run the following command:

pip install sportsfeatures

or install using this local repository:

python setup.py install --old-and-unmanageable

Usage example :eyes:

The use of sportsfeatures is entirely through code due to it being a library. It attempts to hide most of its complexity from the user, so it only has a few functions of relevance in its outward API.

Generating Features

To generate features:

import datetime

import pandas as pd

from sportsfeatures.process import process
from sportsfeatures.identifier import Identifier
from sportsfeatures.entity_type import EntityType

df = ... # Your sports data
identifiers = [
    Identifier(EntityType.TEAM, "teams/0/id", ["teams/0/kicks"], "teams/0"),
    Identifier(EntityType.TEAM, "teams/1/id", ["teams/1/kicks"], "teams/1"),
]
df = process(df, identifiers, [datetime.timedelta(days=365), None], "dt")

This will produce a dataframe that contains the new sports related features.

License :memo:

The project is available under the MIT License.