Data Analysis of Throw-ins in Czech Football League

David Rozlivek

Bachelor's thesis

Abstract:

The increased significance of throw-ins at top-league clubs has recently been seen in practice, as foreign clubs have begun to recruit trainers who are only focused on throw-ins. Thus, this thesis will analyse throw-ins in general, especially how each attribute (e.g., throw-in distance, timing of taking a throw-in, throw-in angle, etc.) influences the completeness of a throw-in, possession retention after a throw-in, and scoring probability following a throw-in. Throw-ins are analysed using the exploratory data analysis and the construction of a basic machine learning model. The examined dataset contains approximately 80,000 throw-ins taken in the Czech Football League from 2015 to 2021. Data analysis indicates that the faster throw-ins are more likely to be successful and possession held. The shorter throw-ins are more likely to be successful and possession retained. Backward throw-ins have bigger probability of being successful and possession kept in comparison to forward throw- ins. A throw-in with a switch, which is taken in defensive or middle zone, has a better likelihood of scoring a goal and also avoiding a conceded goal.

Keywords

Data Analysis, Decision Tree, Set pieces, Throw-ins, Completion, Possession retention, Expected Goals, Machine Learning, SHapley Additive exPlanation, LightGBM model, Football

Language used: English

Date on which the thesis was submitted / produced: 15. 3. 2022

Thesis defence

• Supervisor: Ing. Tomáš Hlavsa, Ph.D.

Czech University of Life Sciences Prague

Faculty of Economics and Management

Bachelor programme:
Business Administration