Data-Driven detection of pitch occupancy in football.

In this thesis, we propose the application of model commonly used in the prediction of human mobility, within a football context. The goal is to be able to predict the future occupation of the football field, starting from information on the previous arrangements of the two teams. The human mobility prediction model chosen was ST-ResNet, a well-known model that has now become a baseline in this art. The ultimate goal is to provide a tool that can help coaches and analysts identify future dangers, or opportunities for their team. In addition, we also develop a software for the study of teams and how they react to any specific situations.