Social Media-Scouting Algorithm Spots Subgroups to Sell Them Stuff


Terry Richardson, the ultimate hipster.

Plaid shirt. Ironic tattoo. Warby Parkers. Outdated mustache. A one-bedroom in Williamsburg. What makes a hipster?

Thankfully, we no longer have to spend our time debating. 

Researchers at the University of California, San Diego, are building an algorithm that will be able to identify hipsters, and a selection of other subgroups, just by using images on social media. They hope that advertisers will use the algorithm to better target individuals.

So far, the scientists have identified several subgroups: bikers, punks, hipsters, goths, surfers and those they call “formal.” Utilizing image recognition software that goes beyond identifying faces, the technology looks at the clothes you wear, your tattoos, your haircut and even your jewelry and places it within an “urban tribe.”

Crazy! Click here for all the math.

Crazy! Click here for all the math.

Then, this information will be used to recommend different products or activities that match your lifestyle. For instance, if you’re considered a hipster, the algorithm might suggest a new brand of organic foods.

So far the technology is far from perfect, with just 48 percent accuracy. Perhaps stereotyping strangers is more of an art than a science.