An infographic titled "How Different Groups Spend Their Day," produced by the New York Times, recently circulated around the Internet, earning acclaim from casual news-readers and the multitude of Twitterers who passed the link around the net. People claimed they liked the infographic, which depicted how different population groups spend their day, because it was interactive, deep, multi-dimensional, and fun to investigate. People also spoke about the results of the quantitative information depicted: claiming that the results were fascinating, identifying trends and specific instances, and even noting some of the peculiarities (the unemployed spend less than 1% of the day working!).
There are at least two aspects of this New York Times infographic I'd like to explore, neither of which relate to the actual data itself. The first question I'm interested in is what it is about the graphic that, despite its flaws, people found so interesting? This relates to a larger question we've been exploring about "cool" infographics and chartporn—depections of data in which people are more interested in the graphic design than its utility. The second question criticises the graphic, examining what could have been done to actually make it live up to the commendations of its blog-sharers, Twitters, and Digg-ers.
I should begin by saying that there are certain structures on the web that lead to the kind of reaction garnered by the "How Different Groups Spend Their Day" infographic. The first is that rarely do people share links to things they dislike. When I saw the graphic, I thought, "this is dumb," but had no inclination to tweet about it because I was indifferent. Stories rarely get shared on Digg, Delicious, Reddit, and other "social bookmarking" websites if people disagree with them (unless the points of the story is to be outraged at some injustice). It should also be noted that the "retweet" function of Twitter encourages people to share links verbatim without applying their own comments or value judgements, which means that people are effectively saying, "so-and-so found this link to be awesome and by sharing it with you I'm agreeing." Think of these points as a kind of disclaimer/criticism of sharing information via the web.
As someone who spent the summer researching infographics for the Newsgames project, I've grown highly critical of most interactive and static infographics I come across on the web (price of doing business, I suppose). But of course, there has to be something that people found interesting such that they wanted to share this infographic. Perhaps the easiest place to begin is at the graphic's opening: what do you do when you open the webpage?
While its use is not directed by instructions, the categories of the graphic invite prompting—everyone reading fits into at least some part of the demographic survey. I'm an employed white male aged 25-64 with an advanced degree and zero children, which gives me five initial places to begin exploring. Cool, done with that. Now, let's see... what would it look like if I were black, or a woman? Hm, interesting. Oh, this seems kind of cool: I can drag my mouse over the graph for more detailed information, or clicks on areas of the chart to isolate them for examination. I can tell there is a lot of data behind this. It covers 18 demographics and 20some-odd categories of daily activities and chronicles them over 10 minute intervals. Perhaps that is what people are attracted to: the multitude of seemingly precise data. That, in itself, would make the graphic worth looking at.
But it was still the experience of using it that people seemed to enjoy most. As I noted above, it's mostly clicking to zoom and hovering to reveal more specific information. If you choose "women" and then click the "TV and Movies" sliver of the graph, the rest of the data disappears and the selected slice moves to the bottom. There is something to be said about the appeal of animation and zooming. When a graph animates, it purports to illustrate some kind of meaningful change—it is not just about points A1 and A2, it's about ΔA. This effect is due to our cognitive associations with scale, and is why something like a bar graph works in the first place. But in the New York Times graph, the zoom animation is meaningless. It merely provides the function of illustrating that the second graph comes directly from the first graph.
The second state of the graph, which isolates a single daily activity within a demographic makes the data more legible. From here a user can click on different demographics to watch the graph change meaningfully. Choosing the activity "work" and exploring by race reveals that whites and Hispanics have similar work schedules, while the black labor schedule is noticeably different. People with no children spend a few minutes more a day eating and drinking, but people with two or more children eat/drink more than those with just one. After seeing these animations, my next inclination is to select multiple categories to see what a retired Hispanic woman spends most of the day doing. Unfortunately, this functionality is conspicuously absent. It would seem that implementing this feature would be a no-brainer (and easy to do because it's just more number crunching). I'd also like to compare side-by-side different demographics in the same category. Alas, I'll have to stick to one at a time.
While exploring the graphic I'm left wondering what it is that I'm supposed to take away from my effort. In the course of our research we've often attributed the proliferation of information visualization as a refusal to embrace one of the basic tenants of journalism: synthesis. The accompanying article says little about the results of the survey, and the textual information displayed while browsing the graph amounts to little more than "fun facts," which may or may not be related to the data being examined. Despite the article describing the results of the survey as "striking," there is no sort of direction that attempts to point to these striking discoveries. Instead, we are left with an information visualization that has enormous quantities of data with relatively few interactions. Perhaps the New York Times could use a tip from children's author Richard Scarry, who posed a more compelling illustration asking What Do People Do All Day?