The recent Ebola outbreak in West Africa, already considered the deadliest in history, led several news organizations to turn to the numbers and data visualizations to tell the story. The Telegraph gets a LAUREL for its mixed map and timeline graphic that also effectively differentiates between five known strains of the virus throughout its history. Though it’s less visually striking than the Telegraph’s, Reuters also gets a LAUREL for its timeline and map hybrid graphic. Less successful was an Al Jazeera map that showed the recent outbreak alongside past ones. While the newer data was mapped down to political subdivisions and included a breakdown of suspected and confirmed cases, data from past outbreaks was only shown at a countrywide level. As a result, the map was trying to compare too many incompatible things. Instead of trying to map everything at the same time, it would have been better to separate out the historical data in a table or timeline and only map the recent outbreak. For ineffectively trying to convey too much information, Al Jazeera gets a DART.
On to maps about a more crucial question: Are cats or dogs more popular pets? That depends on what country (or state) you live in. Working with data provided by market research firm Euromonitor, The Washington Post’s Wonkblog mapped the estimated ratio of cats to dogs as pets across the United States and dozens of other countries. In the American South and Latin America, dogs are the companion of choice, while the Midwest, Northeast, and Western Europe prefer felines. While heat maps are an overused visualization technique in data journalism, often seemingly created merely because they’re easy to make, this piece uses maps effectively as a way to show how population density and cultural divides play out in the world of pet ownership. Wonkblog gets a LAUREL, and not merely because we love our pets.
Since starting this feature on the good and bad of data journalism, pieces by FiveThirtyEight’s Nate Silver have often ended up with darts. That may seem harsh, but that’s only because we expect better from the man who popularized this approach to storytelling. This week, he wrote a piece following up an editorial in The New York Times endorsing the legalization of marijuana. In his piece, Silver suggests that people like those sitting on the newspaper’s editorial board have long supported weed legalization. He does this by extrapolating responses from a series of polls on the issue from people who fit the following criteria, whom he assumes are demographically similar to board members: New Yorkers, 45 years old or older, well educated, liberal or moderate, Democratic or independent, and with incomes of at least $75,000 per year. (He writes: “Perhaps a couple of them live in Connecticut, or one or two of them are younger than 45, or there’s a Rockefeller Republican thrown in, but it should peg them well on the whole.”) If 77 percent of such people believe that marijuana should be legal, then the board probably has the same beliefs too, according to Silver. But the point of polling large numbers of people is not to predict how a small group of individuals may fall on a given issue. The point of polling large numbers of people is to make sure that the differences in individuals don’t skew data. For using polling numbers in the incorrect way, we have to give Silver a DART once again.
Last week, we noted that we had high hopes for The Washington Post’s Storyline as an explainer and data journalism destination for its intention to approach topics in a continual, multifaceted way. This week, Storyline started to take on the issue of bullying using a three-pronged approach: producing a story about real-life bullying at a day camp, crowdsourcing bullying examples, and displaying US Department of Education data that shows how bullying differs by gender and school type. The data component is pretty simple, showing that girls deal with bullying much more than boys and that bullying peaks in middle school. The accompanying story acknowledges there may be some problems with underreporting. The use of data was simple and elegant, but the more impressive thing about it is how it serves as a single component in Storyline’s bullying narrative rather than claim data tells the whole story. For combining storytelling types that include smart data journalism, Storyline gets a LAUREL.