Main Findings
Goals positively affect market value
Assists positively affect market value
Expected goals are important predictors
Age influences player valuation
Detailed football statistics improve prediction performance
2026-06-04
ECON 465: Introduction to Data Science
Final Project Presentation
Gorkem Ozturk
Research Question:
Which player characteristics and performance statistics are associated with football players’ market values?
Why is this important?
Clubs spend millions on transfers
Market values affect transfer decisions
Data-driven valuation can improve decision making
Dataset: Football Player Statistics and Market Values
Outcome Variable:
Main Predictors:
Goals (GLS)
Assists (AST)
Expected Goals (xG)
Expected Assists (xA)
Shots on Target (SOT)
Age
Regression Outcome: Bonservis
Measures player market value
Continuous numeric variable
Suitable for regression analysis
Objective:
Predict player market values using football performance statistics.
Initial Findings
Market value distribution was highly right-skewed
Most players had relatively low values
A small number of players had extremely high values
Solution
Applied log transformation
Distribution became closer to normal
Variables
Goals
Assists
Expected Goals
Age
Purpose
Estimate player market value using basic performance indicators.
Additional Variables
Expected Assists
Total Shots
Shots on Target
Why Build a Second Model?
To determine whether additional football statistics improve prediction accuracy.
| Model | Description |
|---|---|
| Model 1 | Basic Regression |
| Model 2 | Expanded Regression |
Model 2 performed better because it included more detailed player statistics.
Goals positively affect market value
Assists positively affect market value
Expected goals are important predictors
Age influences player valuation
Detailed football statistics improve prediction performance
Single dataset
Limited number of predictors
Market values may change over time
Include multiple seasons
Include contract information
Include wage and transfer fee data
Football player market values can be predicted using performance statistics.
Regression analysis provided meaningful results.
The expanded model outperformed the basic model.
Data science methods can support football valuation and transfer decisions.
Questions?