Visualization of Hierarchical Software Structures

Explorative Analysisand Visualization of

Large Information Spaces

Visualization of Hierarchical Software StructuresMichael Balzer

1 Voronoi Treemaps

Treemaps are a well-known method for the visualization of attributedhierarchical data. Previously proposed Treemap layout algorithms arelimited to rectangular shapes, which causes problems with the aspectratio of the rectangles as well as with identifying the visualizedhierarchical structure. The approach of Voronoi Treemaps eliminatesthese problems through enabling subdivisions of and in polygons.Additionally, this allows for creating Treemap visualizations withinareas of arbitrary shape, such as triangles and circles, therebyenabling a more flexible adaptation of Treemaps for a wider range ofapplications.

2 Treemap Enhanced Graphs

The clear and meaningful visualization of relations between softwareentities is an invaluable tool for the comprehension, evaluation, andreengineering of the structure of existing software systems. Thisapproach presents an interaction and representation scheme for thevisualization and exploration of complex hierarchical graphs toanalyze relations within software systems. Thereby aggregated partsof the software system are represented as Treemaps that visualizethe structure of the contained software entities. An adaption ofexisting rectangle-based treemap algorithms for layouts within convexpolygonal bounding geometries is introduced to allow for adifferentiation of various entity types in the graph visualization.Furthermore, a visual clustering method based on implicit surfaces ispresented to create meaningful visualizations of distorted hierarchicalgraphs of software systems.

3 Clustered Graph Visualization

This level-of-detail technique enables a comprehensible interactivevisualization of large and complex clustered graph layouts either in 2Dor 3D. Implicit surfaces are used for the visually simplifiedrepresentation of vertex clusters, and so-called edge bundles areformed for the simplification of edges. Additionally, dedicatedtransition techniques are provided for continuously adaptive andadjustable views of graphs that range from very abstract to verydetailed representations.

References

[1] Michael Balzer and Oliver Deussen. Exploring relations withinsoftware systems using treemap enhanced hierarchical graphs. InStephane Ducasse, Michele Lanza, Andrian Marcus, Jonathan I.Maletic, and Margaret-Anne D. Storey, editors, Proceedings of the3rd IEEE International Workshop on Visualizing Software forUnderstanding and Analysis, pages 89–94, Budapest, Hungary,September 2005. IEEE Computer Society.

[2] Michael Balzer and Oliver Deussen. Voronoi treemaps. In John T.Stasko and Matthew Ward, editors, Proceedings of the IEEESymposium on Information Visualization, pages 49–56,Minneapolis, MN, USA, October 2005. IEEE Computer Society.

[3] Michael Balzer and Oliver Deussen. Level-of-detail visualization ofclustered graph layouts. In Seok-Hee Hong and Kwan-Liu Ma,editors, Proceedings of the Asia-Pacific Symposium onVisualisation, pages 133–140, Sydney, Australia, February 2007.IEEE Computer Society.

DFG Colloquium Michael Balzer — PhD Track — Associated Member since 01.09.2004Konstanz Work Group — Computer Graphics and Media Design

26 June, 2008 Research Training Group 1042 (GK) — Explorative Analysis and Visualization of Large Information Spaces