David McCandless worked for many years as a writer, but now he likes to call himself a data journalist and information designer. McCandless has never had any formal design training, which makes it all the more interesting that his Web site, Information is Beautiful, has become an influential destination for people interested in data visualizations. He recently published a book of his work, and also contributes to The Guardian’s Datablog.
McCandless marries raw data with design to create images that tell a story. And just as he—and the researcher he works with—puts a lot of thought and effort into the creation of his work, McCandless has also spent time creating mechanisms that enable him to update and correct his designs, share his source material, and encourage people to help him gather and verify data.
Ultimately, he says, the goal is to keep improving the work.
“It’s not just about correcting errors, but it’s also about enhancing and adding and building on [the images],” he says. “The better images are seen as living images, so they evolve over time.”
Aside from the striking visuals involved, the image is notable for the metadata McCandless includes at the bottom. Right next to his name is the kind of version number you’re used to seeing on a piece of software, “1.2.” Over to the right is a list of sources for the information, which includes several newspapers. The text also directs readers/viewers to his site for more information about the data. Curious readers can go to the site, pull up the post with the image, and follow a link to a Google Docs spreadsheet that itemizes all of the data he used and spells out where it came from.
One of the challenges of information design and data visualization is that there are two distinct areas where you can go wrong, which in turn can lead to a range of other errors. First, there’s the data itself. Is it the best available? Are the sources trustworthy? Data is the foundation, so a true visual journalist needs to apply a high level of verification to this element, or the visual display will only serve to call attention to his mistakes.
“As a general policy, we try to use online sources or citations because we’re sharing all the workings and sharing all research in Google Docs,” McCandless says. “We’ve kind of got this first-level rule that we have online links to all the resources and data sources. If we can’t do that, we get on the phone and we quote sources. It’s not ideal, but you can’t always find stuff online.”
And you can’t always get it right. McCandless recently issued a correction and apology for an image he produced about the CO2 emissions of the Icelandic volcano Eyjafjallajoekull.
“It’s the first time we’ve made a mistake on that scale and it was very upsetting, to be honest,” he says. “It’s not great getting things wrong—I hate it. But the really annoying thing is that the image went so far [around the Internet].”
Aside from the data, the design and presentation are the other major place where an image can go wrong. The data needs to be brought to life, but it also needs to be respected. Perhaps an image will be more pleasing to the eye with a nip here, a tuck there, or a slight skewing of the elements. The question is when those cheats become outright distortions.
“One of the unexpected things I’ve gotten [from all of the feedback on the site] is that audiences feel that data is sacrosanct and sacred,” McCandless says. “They won’t accept you doing journalistic tricks or adding little bit of fuzziness with the design or the headline to make it more interesting. They feel data is rigid and should never be played with.”