Football Match Prediction System

Description

Machine learning system predicts English Second Division match outcomes using Logistic Regression on historical data. Features data preprocessing, feature engineering (goals, form), and visualizations (bar/pie charts). Streamlit app offers interactive predictions. Built with Python, Pandas, Scikit-learn, Matplotlib, Seaborn, Streamlit.

Features

• Data Preprocessing: Cleans "England 2 CSV.csv" by handling missing values and filtering consistent teams.
• Feature Engineering: Derives features like average goals (Avg_Home_Goals, Avg_Away_Goals) and recent form (Home_Form, Away_Form).
• Prediction Model: Logistic Regression predicts match outcomes (home win, draw, away win) with probability scores.
• Visualizations:
  • Bar charts for half-time goals, fouls, corners, and yellow/red cards.
  • Pie charts for historical win/draw/loss distributions.
• Streamlit App: Interactive web interface for team selection, stat adjustments, and visualized predictions.
• User-Friendly Design: Prevents same-team selections and provides intuitive outputs.

Technologies Used

• Python 3.8+
• Pandas & NumPy (data manipulation)
• Scikit-learn (machine learning)
• Matplotlib & Seaborn (visualization)
• Streamlit (web app)
• Pickle (model serialization)

Repository Structure

• train_and_predict.py: Processes data, trains model, generates predictions, and creates visualizations.
• app.py: Streamlit app for interactive predictions and visualizations.
• football_prediction_model.pkl: Pre-trained Logistic Regression model (generate via train_and_predict.py).
• README.md: Project documentation.
• Note: England 2 CSV.csv is not included; users must provide their own dataset.

Prerequisites

• Python 3.8 or higher
• A CSV dataset (England 2 CSV.csv) with columns like HomeTeam, AwayTeam, FTH Goals, FTA Goals, HTH Goals, H Fouls, H Corners, H Yellow, H Red, etc.
• Git (optional, for cloning)

Setup Instructions

1. Clone the Repository:

   git clone https://github.com/SATTVIKO/football-match-prediction.git
   cd football-match-prediction

2. Install Dependencies: Create a virtual environment (optional) and install required packages:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install pandas numpy scikit-learn matplotlib seaborn streamlit

3. Prepare the Dataset:

   • Place England 2 CSV.csv in the project root directory.
   • Ensure it matches the expected format (see Technologies Used for required columns).

4. Train the Model: Run the training script to process data, train the model, and save football_prediction_model.pkl:

   python train_and_predict.py

5. Run the Streamlit App: Launch the web application:

   streamlit run app.py

   Access it at http://localhost:8501 in your browser.

Usage

• Training Script (train_and_predict.py):
  • Processes the dataset, trains the Logistic Regression model, and generates predictions for predefined matches (e.g., Blackburn vs. Portsmouth).
  • Outputs visualizations (bar/pie charts) for team stats and saves the model.
• Streamlit App (app.py):
  • Open the app in your browser, select home and away teams, and adjust stats if desired.
  • Click "Predict Match" to view the predicted outcome, probability distribution, and visualizations.
  • Explore historical performance via pie charts and stat comparisons via bar charts.

Example Output

• Prediction:

  Blackburn vs Portsmouth: Home Win (H)
  Probabilities - H: 0.5234, D: 0.2345, A: 0.2421
  

• Visualizations: Bar charts for half-time goals, fouls, corners, cards; pie charts for win/draw/loss distributions.

Future Enhancements

• Integrate real-time data via sports APIs (e.g., Opta).
• Experiment with advanced models (e.g., XGBoost, neural networks).
• Add player-specific features (e.g., top scorer stats).
• Expand to other leagues (e.g., Premier League).
• Deploy to a cloud platform (e.g., Heroku) for public access.

Contributing

Contributions are welcome!

Contact

For questions or feedback, reach out to sattviky@gmail.com

---