How to design charts for the human brain

By Jinghong Chen

At this year’s CAR Conference, Peter Aldhous of BuzzFeed News and Alexandra Kanik, a freelance interactive developer, discussed how to design information graphics for the human brain.

Before visualizing data, Aldhous said, “we should think about how our brains process [the charts].”

In the mid-1980s, renowned statisticians William Cleveland and Robert McGrill did several experiments on how accurately human brains can estimate data presented in different visual cues. They found that the order going from accurate to generic is: length(aligned), length, slope/angle, area/color intensity, volume, color hue.

Aldhous explained that human beings are pretty good on relative positions, but poor on color. This reveals tendencies journalists should be aware of in data visualization work.

Journalists are often making charts to visualize change over time. Bar or column charts are commonly used, yet there are other options like column, line, dot-and-line and dot-column charts.

While the column chart helps people compare individual years, line charts emphasize overall time change. The dot-and-line chart combines these two approaches.

In terms of which type of chart to use, the essential question is: what do you want to show?

There are six possible answers: distribution, relationship, comparison, connection, composition (how parts make up the whole) and location.

Color is also an essential element in data visualization. Thinking about colors in terms of the color wheel is very helpful.

There are three types of color schemes that tell different stories: qualitative, sequential and diverging.

When encoding categorical data, qualitative color schemes are good choices. Sequential colors are more commonly used to encode continuous data. Diverging color is for good for data with positive and negative values.

Whenever you’re using colors, it’s important to make sure that people who are colorblind can still understand the charts. Color Oracle, a free simulator, helps present how people with color vision impairments will see the work.

Aldhous also suggests sketching and experimenting with the data. Try different approaches and show your experiments to people. Then see how they process the data.

Kanik also provided some tips on data visualization:

• Don’t make someone’s brain work harder than it has to.
• Don’t give people too much information at once.
• Human brains can understand motion and sound easily, as they are sequential. Animated gif images are a good approach.

Jinghong Chen is a graduate student at the University of Missouri focusing on data and international journalism.