Document Visualization Overview and Detail by Selectively Varying Density

The challenges of network data display are among the most longstanding in visualization. Current approaches typically rely on some form of network abstraction to reduce network density, or “clutter”, to reveal structural elements salient for the task at hand. Any reduction of complexity must consider the competing goals of providing visual representations that reveal overall, global relations, as well as detailed, local relations. The present work addresses this issue by utilizing the Pathfinder network representation, which allows systematic variation in density, to provide differing, but formally complementary, visual representations. These representations can provide both global context and local detail by selectively increasing structural complexity in a way that overcomes some of the limitations of other techniques. The paper’s examples are drawn from the domain of document relations, though the techniques are broadly applicable.