Integrating Querying and Browsing in Partial Graph Visualizations

Picture of Frank van Ham
Frank van Ham
Published at IBM Technical Report 2011
Teaser image


Partial graph visualizations, sometimes also referred to as online graph visualizations, are visual representations of graphs that show only connections around a particular point of interest. They are often a useful, less complex alternative to visual representations of graphs as a whole or can serve as visualizations when the whole data is not accessible at once. Their biggest downside is that users quickly lose orientation because they never get a sense of the structure of the whole graph. In this paper, we introduce techniques that improve user exploration of partial graphs by integrating both browsing and querying information finding paradigms. From a given partial graph, users can specify queries relevant to user tasks and the results will be displayed as visual aids that helps users navigate representations of graphs with a limited look-ahead. Much like signposts in the physical world, these graphcues highlight the shortest path to relevant nodes from the user's current location in the graph. We demonstrate these concepts on a social network of approximately 500,000 people and 30 million relationships.