Social network analysis is used to find all relationships from the group, dig out the prominent patterns, and observe how information flows between dyads. By social network analysis approaches, users can know how information flows through network ties, how people acquire information and resources, and how cleavages and coalitions operate. In this research, we develop a useful social network analysis system to facilitate teams¡¦ collaboration. The system can draw a social network in ego-centered or whole network layout, and provide information of social network attributes of all users. Both team leaders and general members can make use of it to understand relations and interaction patterns of their team. We also generalize social network attributes to analyze task-based teams at different team development stages for discovering the interaction patterns of different groups in groups¡¦ life cycles. Interaction patterns of members in the team and roles that users play have high influence on a virtual team¡¦s development. With these discoveries, team leaders can obtain concise information about their teams¡¦ performance, and community managers can capture stereotypes of virtual teams in the community. From these evaluation results, we confirm that social network analysis is a useful means to analyze the knowledge activities conducted by virtual teams.