Soccer Player Position Prediction
Applied machine learning to predict soccer player positions based on in-game performance stats using the FIFA dataset.
Player scouting and team composition decisions often rely on subjective judgment. The challenge was to determine whether a player's role on the pitch could be objectively predicted from their statistical performance profile alone.
Built a multi-class classification pipeline using the FIFA player dataset. Engineered relevant features, handled class imbalance, and compared multiple classifiers including Random Forest, Logistic Regression, and Gradient Boosting to identify the best-performing model.
Demonstrated that player positions can be predicted with high accuracy from performance attributes alone — validating a data-driven approach to player analysis that could support scouting and team-building decisions.
A capstone data science project that uses classification models to predict a soccer player's position (e.g., Forward, Midfielder, Defender) from their performance attributes. The project covers data cleaning, feature engineering, model selection, and evaluation across multiple algorithms.