And now, Part Two of the series “Tips for Data Visualization”:

1. Use the Rule of Thirds

What is the rule of thirds?  Imagine taking an image and splitting it into thirds. With this grid in mind, this “rule” identifies the important parts of the image that you should consider placing your points of interest as you frame your image. Take the example above. Which one looks better to you?

Of course you don’t need to pull out a ruler and start measuring your images to make sure they follow this principle, but try seeing the graphic in a way that is elegant in its data-infused beauty.

2. Be consistent when scaling images

If a 1-inch-tall coffee mug represents how much coffee the SSRL staff consumes in one day, how many cups does a picture of a 6-inch coffee mug represent? How can you tell?  If you create your scales with a graphic editor, be careful about being consistent with your images. If possible, use a computer program to help you make consistent measurements.

3.  When using intensity maps – keep the scale coloring from lightest to darkest.

This might seem like a no-brainer, but with the vast range of colors that are available out there, you might be tempted to make your intensity map look like a bowl of Skittles than an image that has real information on it.  Darker colors usually mean higher values, but you can work around this if you define your scale to your audience. Just keep the hues consistent.

4.  Use Maps and Statistics in an Effective Manner

Take a look at the map above.  You can tell the general area where there is great population density, but can you tell which city?  The bubble for Phoenix is swallowing the entire state of Utah here.  Your maps shouldn’t leave your audience guessing.

5. If it looks bad to you, it will probably look bad to someone else too.

As Prof. Jim Lee likes to say, “Use your crap detector!”  Share your graphical information with someone else.  If they take too long to figure out what it means, something is probably wrong.  Make your graphic look good, but understandable.