With the rise of Web 2.0 and collaborative technologies that are attached to,the Web has now become a broad platform of exchanges between users. The majorityof websites is now dedicated to social interactions of their users, or offerstools to develop these interactions. Our work focuses on the understanding of theseexchanges, as well as emerging community structures arising, through a semanticapproach. To meet the needs of web analysts, we analyze these community structuresto identify their essential characteristics as their thematic centers and centralcontributors. Our semantic analysis is mainly based on reference light ontologiesto define several new metrics such as the temporal semantic centrality and thesemantic propagation probability. We employ an online approach to monitor useractivity in real time in our community analysis tool WebTribe. We have implementedand tested our methods on real data from social communication systemson the Web
