In this work we introduce the TimeArcTrees visualization technique for displaying time-varying directed and weighted graphs where the vertices are additionally hierarchically organized. We use a static diagram instead of animation to show the time dimension resulting in graph sequence diagrams that can easily be explored for dynamic patterns such as evolving shortest paths.
To achieve enough space for visually encoding many graphs side by side on screen we first place all vertices of the graph equidistantly on vertical lines. The single weighted and directed edges are represented as curved links, i.e. arcs, where these are categorized into three classes: self-edges, upward edges, and downward edges. Depending on which class they belong to, these are placed either to the left hand side or right hand side of each vertical axis. An additional hierarchical organization of the graph vertices is shown to the left hand side view and can be used to navigate and explore the data on different levels of granularity by expanding or collapsing subhierarchies.
The usefulness of the technique is illustrated by time-varying traffic situation for the German motoway system. The tool is able to compute the shortest path for each graph separately and supports various interactive features.