Dr. rer. nat. Dipl.-Inform. Michael Burch
Email: michael.burch@visus.uni-stuttgart.de


VISUS - Institut für Visualisierung und Interaktive Systeme - Stuttgart

15371818271424282651515113840155351033
Congratulations! 2 In A Row!


Layered TimeRadarTrees

Paper pdf-Version
Layered TimeRadarTrees for distance data from the Soccer Simulation League

Time-varying distance data displayed in a Layered TimeRadarTrees representation. More than 6,000,000 quantitative data points, i.e. directed weighted edges, are displayed in a static diagram without Visual Clutter typically caused by link or node crossings and overlaps.

The Layered TimeRadarTrees technique is an enhancement of the original TimeRadarTrees, that we presented at EuroVis in 2008. The idea is to generate a context as a center wheel that consists of many color coded sectors. The same color coding is used in the thumbnail representations, i.e. miniature representations, that show the outgoing OR incoming edges of each vertex of a time-varying, i.e. dynamic directed graph.

The novel Layered TimeRadarTrees have some benefits compared to their original TimeRadarTrees visualization. First of all, by allowing several thumbnail layers, many more vertices can be represented than in the original approach, i.e. scalability in the vertex dimension is increased. Second, we only display either outgoing OR incoming edges. These can be switched interactively. This leads to a more scalable version for the edges. Third, we do not use a node-link tree diagram on top of the other parts of the visualization. We exploit a radial version of an Indented Pixel Tree Plot, leading to a representation for time-varying weighted directed graphs in information hierarchies without ANY visual clutter caused by link or node crossings or overlaps.

Another improvement of this technique is the fact that dynamic relations between any pair of vertices can be displayed, not only between leaf vertices, as in the original TimeRadarTrees approach, where only dynamic compound graphs have been visualized.

If you are interested in more details about the Layered TimeRadarTrees, please read our IV 2011 paper and/or write an email. Special thanks go to Markus Höferlin for drawing the pictures for the paper and for reading the paper many times.