In the paper we want to argue that semantically informed typed-edge network analysis provide powerful bridge between computational modeling and historical inquiry in the context of prosopography. We focused on a typed-edge network analysis of interpersonal relationships in the court of Emperor Maximilian I: our dataset is a semantic multigraph in which edges represent historically attested actions and are explicitly categorized by interaction type. We apply community detection (Louvain algorithm) to the relative aggregated network and analyze the distribution of interaction types within and across the resulting communities. Our findings show that some communities are semantically coherent, dominated by a specific type of interaction, while others exhibit a broader functional profile. We then define node-level metrics that allow an interpretive classification into different combinations of roles (community broker vs insider, specialist vs generalist).
