Bundling-Aware Graph Drawing Revisited

Area of Science:

• Computer Science
• Data Visualization
• Graph Theory

Background:

• Edge bundling algorithms reduce clutter in dense graph visualizations.
• Current methods often treat bundling as a post-processing step on existing drawings.
• Simultaneously optimizing drawing and bundling offers a new perspective.

Purpose of the Study:

• To investigate a novel algorithmic framework for bundling-aware graph drawing.
• To propose and compare alternative implementations of the Filter-Draw-Bundle approach.
• To evaluate the effectiveness of the new framework against traditional methods.

Main Methods:

• Developed a three-step framework: Filter for a skeleton subgraph, Draw the skeleton, and Bundle remaining edges.
• Proposed and experimentally compared several implementations of this framework.
• Compared against a baseline of drawing the full graph then applying edge bundling.

Main Results:

• The Filter-Draw-Bundle framework demonstrated superior performance.
• Bundled drawings created by the proposed framework outperformed previous approaches.
• Improvements were measured using established metrics for edge bundling and graph drawing.

Conclusions:

• Simultaneous optimization of graph drawing and edge bundling is effective.
• The Filter-Draw-Bundle framework offers significant advantages for dense graph visualization.
• This approach advances the field of bundling-aware graph drawing.