A Framework for an Ego-centered and Time-aware Visualization of Relations in Arbitrary Data Repositories (1009.5183v1)

Abstract: Understanding constellations in large data collections has become a common task. One obstacle a user has to overcome is the internal complexity of these repositories. For example, extracting connected data from a normalized relational database requires knowledge of the table structure which might not be available for the casual user. In this paper we present a visualization framework which presents the collection as a set of entities and relations (on the data level). Using rating functions, we divide large relation networks into small graphs which resemble ego-centered networks. These graphs are connected so the user can browse from one to another. To further assist the user, we present two views which embed information on the evolution of the relations into the graphs. Each view emphasizes another aspect of temporal development. The framework can be adapted to any repository by a flexible data interface and a graph configuration file. We present some first web-based applications including a visualization of the DBLP data set. We use the DBLP visualization to evaluate our approach.