How do you create good visualizations?

November 4, 2008 —

I am sure you have seen graphs that were just horrible to look at. Even worse, it was almost impossible to determine what the graph was trying to communicate. One of the first things you should learn - before you even go ahead and generate a graph - are some basic visualization principles :

Expressiveness The definition of expressiveness reads as follows: A set of facts is expressible in a visual language if the sentences (i.e., the visualizations) in the language express all the facts in the set of data, and only the facts in the data.

Reduce non-data ink This principle has been evangelized quite heavily by Edward Tufte. The data-ink ratio is defined by the amount of ink that is used to display the data in a graph, divided by the total amount of ink that is used to plot the entire graph. This means that, for example bounding boxes, background pictures, grid lines, three-dimensional charts, etc. should be eliminated and simplified.

Emphasize exceptions For example, use the color red to highlight important or exceptional areas in your graph. This makes it easy for a viewer to quickly identify the important information in a graph.

A graph showing how three dimensions are hardly ever a good solution. Occlusion prevents the viewer from seeing the cylinders in the back. The bounding box is unnecessary, the three-dimensional cylinders should be two dimensions and a line chart should probably have been used.

These three principle should help generate nicer and easier to read graphs. You can find more of these simple principles in chapter 1 of Applied Security Visualization. A fun Web site that interactively goes through a graph design IQ test has been built by Stephen Few of Perceptual Edge. It teaches you some of these principles as well.