It’s symptomatic of something that most of the regular comments to this blog now seem to come to the version that syndicates on my Facebook page. I mention this because in response to yesterday’s post on the worthy Chinfographics blog, I received this comment from an old and sharp-eyed friend, Bob:
[T]he front page graphic at the time I’m writing this is a big yellow circle with two much smaller circles below it. It’s supposed to represent the population of China (1.3B) vs. the population of Beijing (12M) vs. the population of Dalian (2.3M) vs. the population of Qiqihar (1M).
Problem is, the proportional difference in the cities’ populations is represented by the DIAMETER of the circle, not the area of the circle. Take a look. Beijing’s circle is not 1/100th the area of China’s, it’s 1/100th the diameter. Likewise with Dalian’s circle & Qiqihar’s circle. So the visual representation is that Beijing is 0.01% of China’s population, rather than 1%.
This is what makes visual representations powerful, of course: what *I* mean by the visual representation of the data may not be what you interpret. Or it can be precisely what I WANT you to interpret.
There are lies, damn lies, and statistics; and then there are visual representations of data. Come to think of it, PR firms and departments should really get behind this…”
So, a few things. First, the graphic in question was in an objective sense wrong, applying a linear formula to a visual representation that was based on area. This was pointed out by a commenter on the Chinfographics site, and to their credit the guys have responded and are addressing it.
Second, Bob, makes a good point. What an author means to communicate through a visualization may not necessarily be the same as how an audience perceives it. Sometimes they can be confusing. More insidiously, because of their power to communicate complex data in very simple ways, visualizations can be also be used to intentionally distort information.
This is not a problem unique to data visualizations. As anyone who followed the Chinese response to CNN and BBC photographs of the Tibet unrest of a couple of years ago, the same thing can happen with the selection and cropping of photographs or video, or even in the editing and presentation of text. How often has a joke in an IM, e-mail or, ahem, blog comment been misunderstood because of missing context that was obvious to the author, but not to the recipient?
But the same storytelling power that makes data visualization so powerful when used well makes them dangerous when inaccurate or distorted. That argues for caution and thoroughness. In the words of that great sage, Uncle Ben Parker, with great power comes great responsibility.
Also, for what’s it worth, in addition to data visualizations there are very good storytelling infographics that are not data based. One now classic (though slightly controversial) example is designer Yang Liu’s “East vs. West” series showing differences in Asian and European culture. Another example is a Men’s Health article that explains the caloric impact of some beverages by showing them next to collections of other junk foods with the same calorie count. There are also various visualizations and infographics for the great Gulf oil spill. These include a Google Earth plugin that superimposes the spill’s area (an admittedly incomplete representation of a three-dimensional catastrophe) over various urban areas, and a complex infographic (large image) on the spill from a company that creates editorial infographics .