In the early 1990s, computer scientist Ben Shneiderman developed a novel form of visualization called a "treemap." Treemaps display hierarchical relationships with nested rectangles. The original motivation for developing this type of information display was to better represent the storage space on hard disks—to visualize the size and organization of thousands of files in multi-level directories and subdirectories.
For UNAIDS (the Joint United Nations Programme on HIV/AIDS), Michael Lindsay of studiovertex designed a pair of treemaps that detail the prevalence of HIV worldwide. The first treemap (left side of the above) visualizes the status of those living with HIV (new infections vs. fatalities; those receiving treatment vs. those waiting for treatment). The second treemap (right side) depicts the geographic regions where those with HIV live, drawing attention to the disproportionate incidence of HIV in Sub-Saharan Africa and Asia as compared to the United States and Western Europe.
MORE ON TREEMAPS
To create a treemap, a tiling algorithm must be defined—a formula and method for dividing a large rectangle into smaller specified areas. The algorithm used in the diagram at right translates the nodes of a tree into horizontal or vertical partitions. Below is an early treemap generated by Ben Shneiderman, showing the file storage on a 1990 color Macintosh.
In addition to the rectangular treemaps above, computer scientists have explored visually representing treemaps using alternate forms
For example, below is a pebble treemap generated by Kai Wetzel that uses nested circles to represent disk usage and the file directory structure. Color denotes the age of files (red=newer, yellow=older). Unfortunately, as Kai notes, the circular forms are spatially inefficient.
Also, another example is the Voronoi treemap. The treemap pictured here was created by Oliver Deussen and uses polygons to display a directory structure of 62,000 files and 42 gigabytes. Brighter colors indicate lower hierarchy.