Multivariate Network Exploration with JauntyNets

The amount of data produced in the world every day implies a huge challenge in understanding and extracting knowledge from it. Much of this data is of relational nature, such as social networks, metabolic pathways, or links between software components. Traditionally, those networks are represented as node-link diagrams or matrix representations. They help us to understand the structure (topology) of the relational data. However in many real world data sets, additional (often multidimensional) attributes are attached to the network elements. One challenge is to show these attributes in context of the underlying network topology in order to support the user in further analyses. In this paper, we present a novel approach that extends traditional force-based graph layouts to create an attribute-driven layout. In addition, our prototype implementation supports interactive exploration by introducing clustering and multidimensional scaling into the analysis process.