Extending NodeTrix with Multivariate Data

Abstract

NodeTrix has been introduced as a visualisation technique for net- work graphs that combines the node-link visualisation with adjacency matrices. This study addresses the gap in current research with regard to encoding multivariate data in NodeTrix visualisations, focussing on edge attributes. This study explores the design space for multivariate encoding of nodes and edges, resulting in a JavaScript implementation of a selection of these encodings for which use cases have been con- structed. This has resulted in design guidelines for those wishing to extend NodeTrix with binary, categorical and numerical data. In doing so, visualisations are enriched with a fuller scope of the data, showing not only what nodes are connected, but also how they are connected. This allows for greater levels of understanding for users, as well as a greater level of storytelling at the side of the designer.