Big Data Goes to the Dogs
I'm fascinated by this graphic on dogs in Slate, which divides dog breeds into four categories: "rightly ignored," "inexplicably overrated," "overlooked treasures" and "hot dog!" Why am I fascinated? Because the graphic is taken from a new book -- sorry, an "infographic mega-tome." And I'm not sure why the author chose to use this precious space to ensure that a large percentage of dog owners will never buy his book.
You will call me oversensitive, because my own beloved bull mastiff is ranked in the "inexplicably overrated" quadrant. Perhaps that's so, but it's hardly unusual. Dog owners are, er, ferociously attached to their breeds. Suggesting that they're idiots who made a bad choice is not a ticket to top-notch sales. Meanwhile, even the folks whose breed ranked in one of the desirable quadrants are sniffily wondering why Rover didn't do as well as the Welsh springer spaniel or the border collie.
But that's not the whole problem with the graphic. The real problem is what even border collie owners will recognize when they look at this graphic: Its ranking is basically useless to anyone who's thinking about getting a dog.
The four-quadrant graph is a classic consulting gimmick; one reason it's so widely used is that it's a really simple way to present data to a large group without delving into tedious explanations. The problem with the four-quadrant graph is that it's really simple, which is tricky if you're trying to describe anything more complicated than a basic accounting metric such as revenue or cash flow.
All that's fair enough -- the short life of a bull mastiff is a major drawback for those of us who love the breed, and I can't say that it didn't come up when we looked to get another dog. But here's the thing: Even if we had decided to look at other breeds, I would never have considered a border collie. Collies are incredibly intelligent, and they're great dogs. They're also a hell of a lot of work, because intelligent dogs are easily bored, and shepherds tend to have a lot of physical energy that needs to be worked off. That's why a lot of city owners fantasize about having their very own Lassie, then drop that idea like a fetched stick as soon as they find out what that would actually entail.
Bull mastiffs are not the smartest dogs in the kennel. Even for the breed, ours has -- and I quote the neurologist we consulted for a congenital condition -- "an unusually small brain." Because we don't need him to herd sheep or pull little Timmy out of the old well, that matters not at all. What matters is that he has a wonderful temperament: affectionate, loyal, alert, gentle, relaxed. Maybe a little stubborn, and definitely a little quirky, but 100 percent charm. He doesn't need to go for a four-mile hike every single day to keep him from destroying our house; mostly, he wants to snuggle up next to us as often as possible. The bull mastiff temperament is what keeps us coming back to the breed despite the vet bills and the 34-pound bag of dog food he plows through every month.
The problem is, it's hard to measure a dog's temperament, which consists of many different traits: from loyalty and protectiveness to sociability with other dogs, from anxiety levels to tolerance for alone time. It's also hard to measure the need for exercise (and for some people, that's a plus rather than a minus anyway). So you end up creating a graphic that tells you everything you need to know about a dog breed -- unless you're considering adopting a member of that breed. In which case, you will throw away this useless graphic and consult less quantitative sources to evaluate whether the dog fits your life and personality.
Graphics can be a great way to easily comprehend information in a glance. But only if the information you really need can easily be compressed into a number. For all the strides we have made in Big Data, there is still a lot of valuable information between the numbers.
*At least, I think that's how he does it; the formula at the bottom of the graph seems to imply that he's counting "ailments" as a positive, but it's not too clear.
This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
To contact the author on this story:
Megan McArdle at firstname.lastname@example.org
To contact the editor on this story:
Brooke Sample at email@example.com