Modelování síly beachvolejbalových týmů na základě dat z olympijských her

This bachelor's thesis focuses on scraping, preprocessing, engineering, and modelling of point-by-point data from the men's beach volleyball tournament at the Olympic Games. To analyse point data from all matches, exploratory analysis methods, rating models, and regression approaches are employed. A key contribution is designing and implementing a point-by-point variant of the Elo model with a dynamic K-factor, which enables tracking team performance over time. In addition, logistic regression models are used to explain the probability of winning a point based on team presence, score development, and match context. The thesis also addresses data quality limitations and team order changes throughout the tournament. Finally, the two modelling strategies are compared in terms of their explanatory power, interpretability, and practical applicability in the field of sports analytics, and potential extensions are discussed.