Quantifying contextual influences on spatial community structure in FC Barcelona’s passing networks

Understanding how contextual factors are associated with the spatial organization of football passing remains a key challenge in performance analysis. While network approaches reveal tactical structures, their interpretability is often limited by a lack of statistical validation and unclear links to match context. This study bridges these gaps by first establishing the statistical robustness of spatial passing communities and then quantifying their association with contextual variables. We analyzed 532 FC Barcelona matches, constructing directed passing networks on a discretized pitch. Community structures were detected using the Infomap algorithm and validated against null models and the Directed Degree-Corrected Stochastic Block Model. The association between communities and contextual factors was measured using Normalized Mutual Information (NMI). Our results, validated by permutation tests, demonstrate that Barcelona’s own coach exhibits the most robust statistical associations with passing community structure, while opponent coach shows only marginal significance. In contrast, match venue and scoreline showed negligible association. Barcelona’s networks consistently exhibited high connectivity, clustering, and a tendency to form two dominant spatial modules, reflecting a stable, possession-based core adaptable to specific threats. This work provides a scalable, context-aware framework for football analytics, advancing from descriptive mapping to diagnostic modeling. Metrics like clustering coefficient and module size offer structure-sensitive tools for tactical evaluation, linking network architecture to actionable insights.