Sequence mining and visual analytics to support tactical analysis in team sports through Process Mining

Data analytics have completely revolutionized professional sports in so many different aspects.The increasing volume of data generated together with data management tools and analytical methods proposed a new data-driven paradigm where players, clubs, agents, and researchers rely on this data to perform tactical, performance or management-based decisions and diagnoses. The main advantage of data is the ability to reconstruct the sports game precisely. Event data is a prominent data source of information in invasion sports games such as football, basketball or handball and consists of a detailed and ordered sequence of all actions that occurred during a match. Despite the highly granular information that event data provides, most of the current approaches fail to properly manage and model the true nature of this data source as a dynamic behaviour. This research project focuses on the applicability of Process Mining to reveal the true power of event data. Process mining is a highly applicable research discipline in the intersection of data mining and business process management. A truly process-aware analysis of event data would acknowledge the existence of actors, roles, resources and sequences in the reconstruction of the game while opening a full set of possibilities for sequence analysis, behaviour analytics and visual understanding of the in-game decision making. By moving away from process-agnostic approaches, this project aims to evaluate event data sources from the main invasion sports, present process-aware data modelling and showcase its main benefits by analyzing sequence patterns and presenting new visual tools.