In the last years, the analysis of data in sports has received considerable attention, especially due to the wide availability of unobtrusive wearable sensors. While most approaches focus on the (post-hoc) monitoring of individuals, a big and still largely unsolved challenge is the monitoring of the tactical behavior and tactical compliance of entire teams in real-time. In this paper, we introduce StreamTeam, a novel and extensible workflow-based approach to analyze data streams and to detect complex team events in real-time. We show the application of StreamTeam to data sets coming from sensors attached to players of football teams.