Tree maps are a way to visualize data that has a hierarchical structure, using rectangles where the size of the surface represents the importance of the respective data element. In some ways, tree maps are similar to a pie chart: If your revenue is broken down into five core product categories, you can use either a pie chart or a tree map to visually show the composition of the various elements. The beauty of tree maps is that you can actually show multiple layers: Let’s say product category A represents the biggest share of your revenue with 31%. Drawing a rectangle of that size shows the overall importance of product category A. But you can then actually break that rectangle down to the next level, and introduce sub-rectangles (to show, for example, the countries where product A is sold). This is something you can’t do in a pie chart.
Tree maps apparently were invented by a University of Maryland professor in the early 90s, who wanted to show visually how the hard disk on the university’s computer was being used (back in the days when hard disk capacity was always limited).
The translation into neatly arranged tiles is actually not trivial. There are various algorithms to optimize the lay-out. A number of software applications exist, allowing users to take data and translate them into a graph. A quick Google and Wikipedia search will give you a good overview of those tools. I have found that for most business applications, a hand drawn set of PowerPoint boxes will actually do the trick.
A great example of a tree map chart is the overview of the S&P 500 “map of the market” on SmartMoney (link). The app shows the overall market cap, by sector, by individual stock within sectors, as well as additional information and recent gains/losses as you scroll over the various tiles. Quite impressive!
