We identify new market and product opportunities to help companies grow.
Call us: 312.276.5140 Email: info@ericksonresearch.com

Research tip: Three ways to make your charts better

As researchers, charts are the primary vehicle for communicating information gained in studies.  Few of us have the luxury of graphic design departments dedicated to creating masterpieces in Illustrator.  Instead, we need to rely on the charting tools available in excel, PowerPoint, or sometimes, SPSS.

While each of these programs makes charting very easy, they also make producing some dreadfully bad charts very easy.  Follow these three simple guidelines, and your charts will instantly become easier to read and help people better understand the data.

Just say no to 3-D.

The only reason to ever make a 3D chart is if the third dimension displays data.  Since, as far as I am aware, this isn’t possible in commonly used software, your clients should never be subjected to a 3D chart in one of your presentations.

Why are these so bad?  The 3D effects, particularly in pie charts, obscure the data. Consider the following example.  Each pie chart shows the same data, but you have to work a lot harder to see significant findings in the 3D version.

“Bad pie chart”:http://www.ericksonmr.com/images/blogimages/badpie.gif”lign

“Good pie chart” http://www.ericksonmr.com/images/blogimages/goodpie.gif

In the 3D pie chart, it’s very difficult to determine the relative size of the slices for East and West.  In the 2D version, this is much clearer.

Perspective is everything.

Most software that does charting likes to help you out by automatically picking a scale for you based on the data.  It is almost never a good idea to accept this default scale because that scale tends to present data in a misleading way.  If the data covers only a very small range, the chart will visually exaggerate the differences, inviting a misinterpretation of the data.  See the example below.

In the first chart, I allowed the default scale as selected by Excel to remain.  In the second, I show the same data, but change the scale to something more reasonable.

“Bad bar Chart”: http://www.ericksonmr.com/images/blogimages/badbar.gif

“Good bar chart”: http://www.ericksonmr.com/images/blogimages/goodbar.gif

There are two big problems with the first chart.  One, it appears that the values for each of the regions are very different.  Two, it appears that the value for the west is very low.  As you can clearly see when the scale is fixed, neither is the case.

Always label the data.

Including data labels makes your chart much easier to quickly understand.  The data labels also allow you to eliminate a lot of extra lines in the chart, making it easier to read overall and, usually, improving the appearance.

The two charts below show the same data.  Notice how adding the data labels to the bar chart above made it easier to figure out what the values are, as well as allowed me to eliminate the value axis and the gridlines, making for a cleaner presentation.

“Bar with labels” http://www.ericksonmr.com/images/blogimages/labels.gif

Use these three painless tips, and your presentation charts will do a better job communicating the substance of the data. 

About the Author

Ed Erickson is the President of Erickson Research, a Chicago market research consulting firm. You can find Ed on LinkedIn and .

Leave a Reply