Extracting and Utilizing Social Networks from Log Files of Shared Workspaces

Log files of online shared workspaces contain rich information that can be further analyzed. In this paper, log-file information is used to extract object-centric and user-centric social networks. The object-centric social networks are used as a means for assigning concept-based expertise elements to users based on the documents that they created, revised or read. The user-centric social networks are derived from users working on common documents. Weights, called the Cooperation Index, are assigned to links between users in a user-centric social network, which indicates how closely two people have collaborated together, based on their history. We also present a set of tools that was developed to realize our approach.