EgoComp: A node-link-based technique for visual comparison of ego-networks

Abstract

Analysis of ego-networks is a critical research problem when analyzing large-scale social networks, as an ego-network represents the social circle a person actually contacts with. One of the core tasks in ego-network analysis is visual comparison, which includes edge comparison and node comparison. Although various works have been done to support comparing normal networks and ego-networks, intuitive node comparison of two ego-networks is still challenging. In this article, we propose egoComp, an intuitive and expressive visualization technique, to analyze the node difference between two ego-networks. To preserve the latent structure of ego-network and lay emphasis on intuitiveness, our design is node-link-based (radial tree layout) and uses a side-by-side method to compare ego-networks. We design a novel storyflow-like graph layout to reveal the relationship of two ego-networks at the individual node level. Furthermore, three different layout algorithms, including origin, greedy, and assignment algorithms, are proposed to meet different user requirements. We demonstrate the effectiveness of our system through case studies and a user study and then discuss the limitations thoroughly as well as the possible solutions and potential future work.