Author: Jonathan Helfman Principal Research Scientist - Oracle Applications User Experience
A treemap is a data visualization technique for displaying hundreds of times more data than a bar graph or pie chart. "Treemaps" get their name because they display hierarchies (tree structures) using nested two-dimensional areas, similar to geographic maps that show states nested inside countries. Treemaps look like a funny grid of colored rectangles. The areas of the rectangles represent a quantitative data dimension, such as a column of quantities from a database table. Larger rectangles typically represent larger quantities. The colors of the rectangles represent a second data dimension, which may be either quantitative or categorical.
Treemaps provide a big-picture overview of a tremendous amount of data, while providing a mouseover tooltip to show additional details on demand. In the example below, the treemap shows thousands of orders in a supply chain organized by distribution center, customer, and item type. The darkest cells show the latest orders while the largest cells show the most expensive orders. The tooltip in the example shows the details of the latest and most expensive order.
In general, treemaps are useful when you need to view large, hierarchical, multi-quantity data sets. However, treemaps do have a few usability restrictions. Specifically, our research has shown that enterprise users can only perceive up to about 3000 treemap cells, and up to about four levels of hierarchy. Also, while cell color can show data that has both positive and negative values, cell areas can only show positive values.
Some types of data are better suited to other visualization techniques. For example, to show only geographic locations, a geographic map is typically better than a treemap. When spatial locations are associated with hierarchical data, however, you should consider combining a geographic map with a treemap. The Supply Chain Management application in the example below provides an overview of stores that sell a range of different product types, which are supplied by three different suppliers. Store locations are shown on the geographic map. The area of each treemap cell shows the "unit supply" for each store, an aggregate value of products already delivered to that store. The color of each treemap cell shows the "unit demand" with lighter colors indicating stores with higher demand. Treemap cells are grouped by product type and supplier. The largest group is selected, causing its members to be highlighted in the map and their details to be shown in the associated table. While the geographic map is great for showing store locations, the treemap is invaluable for providing an overview of the associated product hierarchy. As the amount of data that you need to see grows, treemaps become increasingly important additions to the standard maps, charts, and graphs we rely on for visualizing and understanding enterprise data.
