Friday, May 13, 2016

"Our reader" won't understand something as complicated as that!

The other day I received a gift from Luís Melgar, one of our infographics/data visualization Masters students at the University of Miami (here's our other program). Luís has just graduated, and you should expect a post about his final project soon, as it's an amazing data-driven story. He gave me a printout of this 1870 treemap.

I came back home in the evening and my 10 year old boy got curious about the treemap. He tried to decipher it on his own for a minute, and he couldn't. Then I explained to him that the segments of each square are proportional to percentages, and pointed out a couple of states to illustrate this principle. This took me around thirty seconds.

After that, we spent nearly half an hour together finding the most interesting cases, the outliers, and figuring out which states didn't exist at the time. My kid said: “Numbers are boring, and designing graphics like this must be boring too, but reading it is a lot of fun.” I disagree with him on the first two statements, but agree on the third.

Now, my point: How many times have we all faced the common complaint “our reader won't understand something as complicated as that”? This is a fallacy I spent quite a lot of pages debunking in The Truthful Art*, and the conversation with my kid should be a good argument for the future: If you believe that a substantial portion of your readers (your “average reader” is a mythical creature) won't understand an unusual graphic form, don't dismiss it outright. That's a self-defeating strategy. It'll lead you to stick just to bar charts, time-series charts, and univariate data maps, when other kinds of graphics may be more illuminating —histograms, strip plots, or slope charts, anyone?

Visualization has a grammar and a vocabulary, which can be taught and learned. Scatter plots were unusual in the media just a decade ago, but today it seems that an increasing amount of non-specialists understand them. Why? Perhaps thanks to some pioneers in the news, who decided to start using them years ago, adding captions explaining how to read them. These designers and journalists assumed that most readers may indeed not be able to read those charts at first, but that they are not stupid, and can learn. That's the key: Respect people's intelligence.

Finally, remind your boss that even the all-too-common time-series line chart was “unusual” and “difficult to understand” in the past, but that didn't stop the journalists this post by Scott Klein talks about. They just added a long caption to the graphic. That's the equivalent of my thirty seconds of explanation.

(*There's another more damaging fallacy that I also addressed in the book: “A reader should be able to understand a graphic in five seconds.” Well, no. It depends. Some graphics need some effort, the same way that it takes time to extract meaningful information from a written story, beyond its headline.)