In a training course, one of the students completed his Ph.D. in statistics on presentation of statistics in charts. He objected to the standard exercise using a pie chart. He insisted they were evil and should never be used. While not quite as passionate about the subject, we agree he raised some excellent points.

Pie charts show only ratios, and poorly at that. It is difficult for the human eye to compare slices if there are lots them and they aren’t adjacent. Pie charts suit only a small set of data values, and have major problems if there are any negative values.

The other point that was very important was to avoid the temptation to use lovely 3D effects on our charts. While this looks nice, it can obscure the numbers we are trying to display. The 3D effect often makes values look closer than they really are.

Also think about the sequence of your values in the chart. Select a sequence that best displays the data.

Some factors to consider for best readability:

• Point of the chart. Ask why this chart exists, and make decisions based on your answer.
   

• Information focus. What information does the user most need? Least need? Decide what's important and leave out everything else. Lasers are powerful because they focus. Do the same, so your charts are powerful.
   

• Color contrast. Because your information is focused, you don't need a large pallet of colors because there aren't many things to color. Aid the eye by using contrasting colors.
   

• Font size. Don't reduce the font as a way of cramming more onto the graphic. Instead, write less. Use short tags and labels, rather than long ones.
   

• Amount of information (lean toward "less is more"). The key points will get lost of there's too much information.
   

• Chart type. Use the chart type that best conveys your central message or best permits a clear portrayal of the data. Most people go straight to the pie chart, even though that's seldom the best chart type.