2.1 Good Visualization Principles

Before we discuss the standard graphics, let’s lay out the basic design principles for good data visualizations.

1. Show the data: This may be self-explanatory, but make sure that the data is the focus and driver of the visualization.

2. Avoid distorting the data: Avoid 3D charts as the added dimension distorts the comparison. The area in a graph should equal the magnitude of the data it is representing.

3. Simplify: In 1983, Edward Tufte said that “A large share of ink on a graphic should present data-information, the ink changing as the data change. Data-ink is the non-erasable core of a graphic, the non-redundant ink arranged in response to variation in the numbers represented.” Remove any unnecessary “ink” that does not assist in the presentation of the data. Remove distractions.

4. Facilitate comparisons: In order to explain variation, we want the graphics to facilitate comparisons between groups. The design should make it easier to compare between groups rather than harder.

5. Use contrast: Humans have developed to seek out visual contrast. When choosing colors and annotation, strive for more contrast in luminance (white to dark) to make it easier for everyone to visually perceive.

6. Use color appropriately: Think about your audience. A small proportion of the population is color-blind; try printing it in grayscale to see if the color palette is still effective. Also, every culture has different associations with colors; ask others for feedback on color choices. Neuroscience research has shown that humans are more sensitive to red and yellow, so those are good colors to use for highlighting key points.

7. Annotate appropriately: Informative text is crucial for providing data context. Make sure to use informative axis labels and titles. It may be worth adding text to explain extreme outliers.

For examples of good data visualizations in the news and discussion around them, check out the New York Times column “What’s Going on in This Graph?”.