Abstract

This paper demonstrates the application of graph drawing and information visualisation techniques to the visualisation of information which can be modelled as an attributed graph. An attributed graph can be used to model a wide range of different types of information, including system descriptions and database content. We propose the novel Composable Layouts and Visual Sets (Clovis) class of views, and describe supporting software component infrastructure, including a user interface for creating and interacting with Clovis views. A framework for composing graph vertex layouts is presented, including the division of responsibilities between the layout strategies being composed and the mechanism for coordinating their execution. Three broad classes of layout strategy are identified, and opportunities for novel hybrid layouts highlighted. The definition of sets of graph elements and the allocation or overlaying of distinctive visual attributes to members of the set are combined under the notion of a visual set. A visual querying mechanism for the allocation of graph elements to visual sets and for the clustering of graph vertices in preparation for layout composition is described. The versatility of the Clovis view family is demonstrated through its application to a variety of problem domains, and future research directions are identified. Keywords: Database visualisation, attributed graph, graph drawing, clustered graph, layout composition, overlay, Tree Map.

Citations