Od podatkov do prvakov: integracija podatkov za podporo odločanja v športu

The use of data has become an important part of managing professional sports teams in recent years. Football has been a bit later in this data evolution than some other sports, mainly due to the very nature of the game, which is dynamic, less structured and involves a large number of players moving simultaneously across a large space. In this paper, we present how raw football data can be used to obtain key information that coaches, analysts and other members of the coaching staff use in their decision-making processes. Coaches and analysts are not just external observers, but are actively involved in the process of creating metrics and models. The central part of the paper presents a practical example in which we translated the coach's principle of defensive compactness into a data model and created metrics that made it possible to monitor whether the team is actually implementing the agreed principles on the field and where deviations occur. We used a combination of the DuckDB and MotherDuck systems and the Python programming language for data processing, transformation and modeling, and presented the results through interactive reviews in Streamlit and automatically generated reports in PDF format. With this article, we want to show that data in football does not replace the coach's knowledge, but rather complements it by making tactical ideas measurable, verifiable, and useful in everyday work.