Predicting Football Player's Salary

Presentation Type

Poster

Student

Yes

Abstract

The increasing financial demands in football necessitate data-driven strategies for structuring player wages. This study aims to develop a predictive model for football players' salaries using statistical and machine learning techniques. Leveraging a dataset of 99 players with 42 attributes sourced from futhead.com via Kaggle and Scrapy, we conducted exploratory data analysis and applied various modeling approaches. Linear regression identified Age and Value as significant predictors of wages. Ridge regression emerged as the most accurate model, yielding the lowest RMSE (58,699) and MAE (48,336). The study highlights challenges, including the exclusion of high-cardinality variables (e.g., Club and Nationality) and limited feature engineering, which constrained model complexity. The findings underscore the influence of player attributes like Age and Value on wages, offering clubs practical tools for budget management and transfer decisions. Future research will focus on incorporating non-linear models and exploring domain-specific features to enhance predictive accuracy.

Start Date

2-7-2025 1:00 PM

End Date

2-7-2025 2:30 PM

This document is currently not available here.

Share

COinS
 
Feb 7th, 1:00 PM Feb 7th, 2:30 PM

Predicting Football Player's Salary

Volstorff A

The increasing financial demands in football necessitate data-driven strategies for structuring player wages. This study aims to develop a predictive model for football players' salaries using statistical and machine learning techniques. Leveraging a dataset of 99 players with 42 attributes sourced from futhead.com via Kaggle and Scrapy, we conducted exploratory data analysis and applied various modeling approaches. Linear regression identified Age and Value as significant predictors of wages. Ridge regression emerged as the most accurate model, yielding the lowest RMSE (58,699) and MAE (48,336). The study highlights challenges, including the exclusion of high-cardinality variables (e.g., Club and Nationality) and limited feature engineering, which constrained model complexity. The findings underscore the influence of player attributes like Age and Value on wages, offering clubs practical tools for budget management and transfer decisions. Future research will focus on incorporating non-linear models and exploring domain-specific features to enhance predictive accuracy.