TagNetLens: Multiscale visualization of knowledge structures in social tags

Social tags reflect personal and shared vocabulary, and provide opportunities for people to organize and search information. However, tags are usually not structured. To find relevant tags and associated documents, people often need to invest significant amount of cognitive resources to make sense of the relationships among tags. To help the sensemaking of social tags and exploration of knowledge structure of them, we propose an approach of tag networks, TagNet, in which tags are linked by their corresponding documents and a multiscale tag hierarchy are derived with network clustering and aggregation techniques. We also present TagNetLens, an interactive tool that allows users to explore a tag network and its tag hierarchy. We report a case study of TagNet and TagNetLens based on social tags and documents from CiteULike. The results indicate that our TagNet approach can provide users with knowledge structures that are similar to cognitive structures of concepts in people's minds, and TagNetLens can help people to better explore the space of social tags and may have potentials to facilitate the understanding of the knowledge structure in social tags.