June 23, 2009
Back to the roots of the treemap Data Visualization
In the 1990's, treemaps were not very well known and most people looked at them as an exotic and unusual type of data visualization. Treemaps let you focus on the elements that are most important to you, make it obvious which elements need immediate attention and help you spot outliers and patterns without hardly thinking about it.
The first thing to understand about treemaps is that these graphs don’t use a traditional rows and columns-oriented matrix layout. They display a data tree rather than a data table with rows and columns. The x- and y-coordinates of the position where a rectangle is located in a treemap doesn’t really mean anything. Instead it is the containment structure of the treemap that carries meaning. Recursively nested rectangles illustrate the hierarchy in the data tree — enclosure of one rectangle inside another denotes that we’re looking at different levels of the tree simultaneously.
The essential novelty of treemaps — and what set them apart from heatmap displays — is that they use area to encode numerical values. The size of a rectangle in a treemap could correspond to some numerical dimension in your data, say sales in US Dollars for a specific type of product. This technique gives more display space to data elements that are important because they have relatively larger values. When this is combined with a good color scheme, treemaps make it immediately obvious which elements are significant and need urgent attention.
Once you have understood the above basic concept of the treemap, including the important fact that it doesn’t use x- and y-axes in the traditional way, you will find that it can help you to immediately spot outliers and quickly find patterns quickly even when it comes to a large numbers of data points.
Today, treemaps are much more popular and are likely to be seen in a wide variety of Business Intelligence applications and dashboards. Treemaps are no longer only used in very specific scenarios and are becoming more of a general purpose graph type.
Let's look at the roots of the treemap for a minute. One of the original applications of the treemap display was to visualize space usage on hierarchical file systems, as described by the inventor professor Ben Shneiderman on this web page:
http://www.cs.umd.edu/hcil/treemap-history/
Here is a link to my personal favorite tool that does just that – Space Sniffer. I find this freeware Windows application to be small, simple, free, fast and fun:
http://www.uderzo.it/main_products/space_sniffer/index.html
Check it out for yourself – run it to monitor space usage on your hard drive. Personally, I think it demonstrates how efficiently treemaps can help you monitor resource usage.
If you want to use treemaps in a resource monitoring application in your business, get an evaluation copy of Panopticon EX and try it for yourself:
http://www.panopticon.com/products/trial.htm
We have clients using treemaps in a very diverse set of industries, including pharmaceuticals, engineering, Consumer Packaged Goods (CPG), heavy manufacturing, telecommunications, retail, and financial services.
Markus Skyttner
CTO
Panopticon Software
More Blog Posts
|
