Constant density visualizations of non-uniform distributions of data

The cartographic Principle of Constant Information Density suggests that the amount of information in an interactive visualization should remain constant as the user pans and zooms. In previous work, we presented a system, VIDA (Visual Information Density Adjuster), which helps users manually construct applications in which overall display density remains constant. In the context of semantic zoom systems, this approach ensures uniformity in the z dimension, but does not extend naturally to ensuring uniformity in the x and y dimensions. In this paper, we present a new approach that automatically creates displays that are uniform in the x, y, and z dimensions. In the new system, users express constraints about visual representations that should appear in the display. The system applies these constraints to subdivisions of the display such that each subdivision meets a target density value. We have implemented our technique in the DataSplash/VIDA database visualization environment. We describe our algorithm, implementation, and the advantages and disadvantages of our approach.