Fans without borders

On Thursday, the internet was aflutter with an interactive map put out by the New York Times that purported to show breakdowns of baseball fandom by zip code across America. "A Map of Baseball Nation," it was called, in grand Times style. It was exactly the kind of flashy, quant-y thing people go gaga for in the post-Nate Silver world. For Mets fans, though, the Times' map brought sobering visual evidence that their team does not "own" any territory at all.

According to this map, the Mets have no home, anywhere. All portions of the tri-state area pledge their allegiance to the Yankees. Within the five boroughs, it's the same story. The Mets can't even claim Queens or pockets of Brooklyn or Staten Island I'd always assumed to be theirs. (Growing up, I was under the impression that everyone on Staten Island was a Mets fan, based on my scientific polling of every relative I knew who lived there.) Long Island is supposed to be Mets Country, but if the map is to be believed, not a single part of Nassau or Suffolk leans toward the Amazins.

There are several multi-team regions in the baseball world—Chicago and LA being the biggest—but the Mets were only one of two teams with local competition that failed to capture any turf. The luckless Oakland A's, whose surprising competitiveness in recent years has been undermined by their literally crappy stadium and the Giants' two World Series rings, were the other.

On its surface, this news is powerful fodder for the LOLMets cannon, and as bad news for the franchise the Times' map can't be completely dismissed. However, while visualizations are fun things to RT and fave on the interwebs, what they can definitively tell you about reality is another story.

For any visualization to have merit, it has to be based on solid data. Very few people who have harped on the map's conclusions have mentioned that those conclusions are drawn from Facebook Likes. They don't mention this because it makes the whole enterprise seem ridiculous. Tell people you've based a supposedly scientific assessment on anything Facebook related and you might as well hand out investment advice based on Magic 8 Ball responses.

What correlation do Facebook Likes have to Real Life? As the Magic 8 Ball might say, "Answer unclear." A study at Cambridge University found that Facebook Likes could be used to accurately determine personality traits. But the same study classified them as a "very generic set of digital records" and concluded that a person's web history or credit card records could yield similar results.

The problem with basing a visualization on Facebook Likes is that the Like has no value as an indicator for level of enthusiasm, for any endeavor. When you click that you "like" something on Facebook, that Like stands in for an enormous gamut of emotions, ranging from I Guess It's Okay all the way up to I Will Die For It.

For people who make TV shows and movies and music, Likes can be helpful indicators of how many people are into their product. Creators don't particularly care how hard you enjoy their creations, so long as you watch/listen. Sports are a different animal. Casual fans of a team, who check in and out of a season as mood dictates, have very little value to a team on levels both ephemeral (will they support us through hard times?) and material (will they spend money at the ballpark?).

The assumption of the Times' map is that everyone who says they "like" a team on Facebook is a die-hard fan of that team, but there is a very big difference between "liking" a team and being a fan of one.

When someone's Facebook profile says they "like" the Yankees, all we know really is that person doesn't hate the Yankees. We don't know if this person could name anyone on the team other than Derek Jeter, or if he/she watches the games with any regularity. "Liking" requires no commitment. It asks nothing of a person but a nanosecond of their time. It's impossible to say how a Like serves as an indicator of a team's "real" fans. This would probably be even more true in one-team regions, where a person might click Like for a team out of pure civic pride while possessing zero affinity for the sport that team plays.

We also have to look at how this map declares its winners. In this regard, it is closer to an electoral vote map than a popular vote one. A team wins all of a zip code by taking a simple majority of the Likes in that zip code. Just like in a political election, how many people didn't vote at all is irrelevant to the result; it's winner take all whether everyone casts a ballot or one lonely guy does. Knowing 40 percent of the Facebook Likes in a certain area favored a certain baseball team doesn't tell us how many people in that area don't care for baseball at all. For all we know, the results could be skewed by a hardcore group of fanatics fulminating amid a sea of baseball apathy.

Basing a map on zip codes is a highly suspect enterprise. There is no standard definition of what a zip code translates to in terms of number of residents or geographical area. All it means is there's a post office somewhere with those numbers on it. In New York City, zip codes represent very small slices of land with dense populations. In large swaths of the country, a zip code could represent a huge expanse of land with relatively few people in it.

Look at the Times' map and you'd think the Colorado Rockies own half the west. All that purple looks impressive, until you consider that big chunks of the territory they "own" include parts of Wyoming and Nebraska that are some the least populated portions of the lower 48. The Rockies claim more land as their own than the Mets, but do they have more fans? A NY1 poll in 2009 estimated 25 percent of the NY population rooted for the Mets (versus 34 for the Yankees and 35 who don't root for anyone local). In New York City alone, that translates to 2 million people, give or take. Denver has a population of a little over 600,000 total.

Go a little west of Rockies Country in Utah. Zoom in on the map and you'll see the Yankees "own" Millard County in that state by virtue of winning 20 percent of its Facebook Likes. The Yanks just edged out the Red Sox at 19 percent. The Braves came in third at 7 percent. What does this tell us about Millard County, Utah, population 12,503? Does this little spot on the map have a large population of ex-New Yorkers battling it out with refugees from Southie, while bemused former Georgians look on? In all likelihood, the number of hardcore baseball fans in that county is so minuscule that they reflect very little of the local enthusiasms as a whole. The map stamps Millard County with pinstripes all the same.

This is not to say that the Times' map has no bad news for the Mets. Though zip codes are a flawed basis for judging levels of fandom, that the Mets can't claim a single one, even in Queens or on Long Island, is upsetting. And while Facebook might be the oldest social medium in terms of the median age of its users, social media still skews young in general. One could draw the conclusion that the Mets' failure to make an impression on Facebook users means they are losing out with the younger generation, which does not bode well for the future of their fan base. 1986 is a long time ago now, and it ain't getting any less long ago. Even 2006 means nothing for a kid who is just now reaching the age where s/he is getting into sports. They have no living memory of a Mets team of any relevance.

For a map to be useful, it has to reflect the actual landscape and show you how to get where you want to go. The Times' map of baseball fandom is incomplete in the former respect. As far as the latter is concerned, hopefully it will impress upon the Mets that they have to get going somewhere, and soon.