Guided navigation in task-oriented 3D graph visualizations

Abstract

In many application areas an optimal visualization of a complex graph depends on the specific task to be accomplished by the user. Therefore a means of locally customizable layouts is required to focus on different problem aspects. We propose a concept of user-driven hierarchy formation and layout optimization in the context of IBM zSeries I/O configurations. We combine the concepts of glyphs, local 3D layout managers and partitioning galaxies to create an optimal layout for a given task. Additionally, to simplify global navigation, structural details are abstracted and can be refined upon user request. Since element context is crucial for understanding and consistently configuring an I/O configuration, guided navigation along the connections in this data structure is essential. To efficiently customize a graph layout, complex selection mechanisms are needed to quickly define the areas affected by a particular layout. Although we focus on IBM zSeries I/O configurations, the approach we present is quite generic and can be adopted for other fields of application.