Predicting Ball Location From Optical Tracking Data

In this study, an automated method for predicting the ball’s location during a soccer match has been developed using optical tracking data. The rolespecific analysis using the individual player attributes has been conducted on a dataset of 300 matches from the Turkish Football Federation Super League 2017-2018 season (≈34,000,000 data points).

The data is provided by an optical tracking system developed by start-up company Sentio Sports Analytics.

The project contains data analysis, features construction, model development and testing files written using python.

This repository is part of our 2022 paper titled: "Prediction of the Ball Location on the 2D Plane in Football Using Optical Tracking Data"

Orange and blue point -> home and away team players, respectively
Green dot -> actual ball location
Red dot -> predicted ball location

License

This library (all the notebooks) is distributed under Apache License 2.0 . Please see Apache License 2.0 terms to learn about how to use this library.

Project Instructions

Getting Started

• Clone the repository, and navigate to the downloaded folder.

git clone https://github.com/anaramirli/predict-soccer-ball-location.git
    cd predict-soccer-ball-location

• Create (and activate) a new environment with Python 3.6 and the numpy package.

* Linux or Mac:

conda create --name my_env python=3.6
    source activate my_env

* Windows:

conda create --name my_env python=3.6
    activate my_env

• Check requiremenets.

requirements.py