Analyzing Dynamic Football Passing Network

Abstract

In this thesis we are concerned with the analysis of football matches represented as passing graphs, where players are nodes and directed edges indicate the passing of the ball. As opposed to previous work, we label the edges of the graph with the time instants when the ball is passed between two players, thus constructing a time-varying graph. We then employ techniques from social network analysis to study centrality roles and other indicators, keeping into account the evolution of the game in time. More precisely, we focus on degree centrality, closeness, betweenness, pagerank, eigenvector centrality, and clustering coefficient, and we compute these measures dividing the overall play time into time windows following two different models. The results are to be compared with a static analysis performed on a unique time window. Our study provides observations that are different and possibly more accurate than the ones that can be obtained by a static analysis, opening the door to a dynamic study of football matches.