Soccer Match Outcome Prediction

Contributors

Team Member	Unity ID
Apurv Choudhari	apchoudh
Chinmay Singhania	csingha
Parth Kulkarni	pnkulka2
Madhur Dixit	mvdixit

Project Overview

This project predicts soccer match outcomes using different datasets and methodologies. Below is an overview of the files in this repository:

Files

1. Match_Outcome_Prediction_Manual_Feature_Engineering.ipynb (Main File)

   • Implements manual feature engineering techniques for predictive modeling.
   • Uses database.sqlite dataset.

2. Match_Outcome_Prediction_Featuretools.ipynb

   • Employs automated feature engineering using Featuretools.
   • Also uses the database.sqlite dataset.

3. Match_Outcome_Prediction_New_Dataset.ipynb

   • Works with a simpler dataset derived from 10 CSV files, containing data for seasons 2015-16 to 2024-25.
   • Combines these files into a single dataset for predictions.

4. README.md

   • Documentation for the project.

Datasets

Our combined data (Contains data from both the sources): Combined Data

1. Kaggle Dataset: European Soccer Database

   • Format: database.sqlite
   • Used in: Match_Outcome_Prediction_Manual_Feature_Engineering.ipynb and Match_Outcome_Prediction_Featuretools.ipynb.

2. Football Data UK: England Matches Dataset

   • Format: 10 CSV files (2015-16 to 2024-25 seasons).
   • Used in: Match_Outcome_Prediction_New_Dataset.ipynb.
   • Text file explaining the meanings of the columns for the dataset. Text file with column meanings

Key Differences

• Manual vs. Automated Feature Engineering:

  • Manual_Feature_Engineering.ipynb applies manual techniques.
  • Featuretools.ipynb uses automated feature extraction.

• Datasets:

  • database.sqlite contains a structured database of European soccer matches.
  • The CSV dataset combines English Premier League match results for 10 seasons.

Installation and Setup

1. Clone the repository.

   git clone <repository-url>
   cd soccer-match-outcome-predictions

2. Install dependencies.

   pip install -r requirements.txt

3. Ensure the datasets are placed in the correct directories:

   • Download the data files from Combined Data.
   • Place database.sqlite in the root folder.
   • Combine the 10 CSV files into one file (combined_dataset.csv) using Match_Outcome_Prediction_New_Dataset.ipynb.

How to Run

1. Launch Jupyter Notebook.

   jupyter notebook

2. Open the main file: Match_Outcome_Prediction_Manual_Feature_Engineering.ipynb.

   • Follow the steps in the notebook to preprocess data, train models, and evaluate predictions.

3. Optionally, explore other notebooks:

   • Match_Outcome_Prediction_Featuretools.ipynb for automated feature engineering.
   • Match_Outcome_Prediction_New_Dataset.ipynb for working with the simpler dataset.

Notes

• The notebooks are designed for educational purposes and might require GPU resources for larger datasets.
• Make sure to have Python 3.8 or later installed.

---