GEEK FIGHT!!! says JD Long on Twitter as the New York Times publishes a widely-reposted interactive graphic about Netflix rental data, and the Wall Street Journal also gets into the interactive-viz game with a graphic on bank bonuses.
If it’s a fight, it’s a knockout in the first round, if you ask me. There’s no surprise why the Netflix chart has been tweeted and blogged to death since it was released: it’s a joy to play with and explore the patterns between DVD titles and their relative popularity between different cities and districts within. It’s no surprise to learn that Milk was wildly popular in the San Francisco area (though comforting to see it confirmed); more interesting is the pattern in a red-state city like Atlanta. Interest in the film in the cosmopolitan city core dwindles rapidly in the suburbs:
You can spend hours exploring the pattens of other movies and asking yourself why the pattern looks as it does (and sometimes the more interesting question is, why doesn’t it?) A good visualization prompts followup questions and gives you the means (or at least the motivation) to answer them. (By the way, the NYT did use R to look at the data using Principal Components Analysis. But the results of PCA are difficult to express in a way that easily be consumed by readers of a graphic like this, so none of the R-based analysis made it into the chart online. Maybe next time.)
Compare that to the WSJ’s effort. The data source isn’t nearly as rich: simply the proportion of revenue allocated to compensation and benefits in various Wall Street firms. With all the furore about record bonuses it’s certainly a topical chart, but for me the delivery leaves me … well … shortchanged. I can’t get over the data presentation — pie charts when presented as squares are still pie charts, and I find it impossible to generate any meaningful comparisons from them. The fact that none of the detail charts are labeled by company name (except by rollover) inhibits meaningful comparisons, too.