PCA Analysis on Football Player Performance
Applying Principal Component Analysis (PCA) and SVD to the FIFA 23 dataset to identify key performance factors, segment player profiles, and generate recruitment insights via K-means clustering.
Business Context
A mathematics project applying dimensionality reduction techniques to a real-world dataset — European football player statistics from FIFA 23 — to uncover hidden performance patterns.
Strategic Problem
How to reduce the high dimensionality of football player statistics to identify the primary factors that differentiate player profiles and support data-driven recruitment strategies?
Data Sources
FIFA 23 Complete Player Dataset from Kaggle — performance metrics for European football players including physical, technical, and tactical attributes.
Methodology
Data preparation (missing values, outlier detection, Z-score standardization). Constructed correlation matrices and determined optimal components. Applied PCA with correlation circle visualization and player projections onto factorial planes. Implemented SVD on normalized matrices and compared with PCA results. Applied K-means clustering on principal components to segment player profiles.
Key Results
Identified primary performance factors explaining player variance. Segmented players into distinct profiles using K-means on PCA components. Generated actionable recruitment and training strategy recommendations.
Business Impact
Demonstrated strong mathematical foundations in statistics and linear algebra applied to real data — from PCA theory to actionable player segmentation insights.