Evaluation of Hierarchical Visualization for Large and Small Hierarchies

Hierarchical visualization can provide a useful abstraction for exploring and analysing complex datasets. Many different methods for visualization of hierarchical data have been developed and determining which is the most suitable for a given dataset can be a challenge. Previous evaluations of hierarchical visualization methods typically focus on comparing hierarchical structures and less so on how well they represent underlying data. In this paper five different visualization methods for hierarchical data are evaluated: Treemaps, Icicle plots, Sunburst, Circular Treemaps and Bubble Treemaps. Each method was compared at two different sizes of hierarchical structure with questions comparing nodes based on their underlying data features, through size and colour. The initial results from this study indicate that for large hierarchies, Sunburst and Circular Treemaps are the best overall performers, while for small hierarchies, Treemap, Icicle plot and Circular Treemaps performed best.