FIFA Match Total Goals Predictor

This project builds a machine learning model to predict the total number of goals scored in European football matches using the FIFA dataset. It draws on two data sources — historical match records and team tactical attributes — which are merged to create a rich feature set covering both teams in each fixture. The pipeline covers the full workflow: data loading from a SQLite database, feature selection and engineering, handling missing values, model training using Linear Regression, and performance evaluation via MSE, RMSE, MAE, and R². Betting odds (Bet365) are included as features alongside tactical metrics such as build-up play speed, chance creation, and defensive pressure for both the home and away sides. Results are visualized through three charts: an Actual vs. Predicted scatter plot, a Residual plot to assess model behavior, and a Feature Coefficient chart highlighting the most influential variables. The project demonstrates a complete end-to-end data science workflow applied to real-world sports data. Tech Stack: Python · Pandas · NumPy · Scikit-learn · Matplotlib · SQLite