A sensemaking approach to visual analytics of attribute-rich social networks.
详细信息
文摘
Social networks have become more complex, in particular considering the fact that elements in social networks are not only abstract topological nodes and links, but contain rich social attributes and reflecting diverse social relationships. For example, in a co-authorship social network in a scientific community, nodes in the social network, which represent authors, have multiple attributes, including academic rank, affiliation, expertise domain, and so on. Analyzing such attribute-rich social networks requires making sense of not only its structural features, but also the relationship between social attributes and network structures. Current social network analysis tools usually treat actors and relationships in social networks as abstract topological components, i.e., nodes and links, and ignore the impacts of social attributes over network structure. Furthermore, these tools often combine social network analysis with statistical methods and produce complex visual results； however, they often overlook users&rsquo； needs for sensemaking tools that help to gather, synthesize, and organize information. To address these challenges, this study proposes a sensemaking approach to visual analytics of attribute-rich social networks in this dissertation. A sensemaking framework is first presented, which consists of both bottom-up processes, constructing new understandings based on collected information, and top-down processes, using prior knowledge to guide information collection, in analyzing attribute-rich social networks. Then, this study introduces a graph model of attribute-rich social networks and designs a novel visualization technique, TreeNetViz, to analyze network patterns over social hierarchical attributes. A case study shows that TreeNetViz help users to understand multiscale and cross-scale network patterns for networks with hierarchical attributes. Further, this study also proposes new methods of actor similarity analysis to leverage both network information and social attributes with the graph model and demonstrate their applications in the field of information retrieval and information visualization. Finally, guided by the framework, a system, SocialNetSense, is designed and developed to support the sensemaking in visual analytics of attribute-rich social networks. Results from a case study of using our system to analyze a scholar collaboration network shows that this approach can help users gain insight into social networks both structurally and socially, and enhance their process awareness in visual analytics tasks.