Analysis and Prediction of Football Players' Market Value

The bachelor thesis focuses on the analysis and prediction of the market value of football players based on their performance and static attributes. It concentrates on players active in the English Premier League over seven seasons (2017/2018-2023/2024), using publicly available databases as primary data sources. The main objective of the thesis is to develop a predictive model that estimates a player's market value at the end of the season using historical data. The model incorporates the player's initial market value, immutable characteristics (e.g., nationality, preferred foot), and seasonal performance metrics. A significant part of the thesis involves the construction of the main dataset used for training and evaluating different types of regression models. The dataset contains a total of 2230 records representing players across multiple seasons and covers data for 861 unique players. The models were evaluated using standard regression metrics (MAE, RMSE, MAPE and R²). The tested models fall into two categories: linear regression models and tree-based models. The tree-based models outperformed the linear ones overall, with the XGBoost Regressor achieving the best results, reaching a MAE of €3.12 million and an R² score of 0.935.