Craigslist's "Missed Connections" Get Matched By An Algorithmic Cupid

There you are, on your morning subway commute just like any other day. You’re reading Murakami’s 1Q84 and glance up periodically as people shuffle in and out of the car. This time, a woman sitting down across from you catches your attention. She gets situated and pulls out the same book you’re reading from her bag. You shoot her a look, and she notices and smiles back. Neither of you say anything, but oh look, this is your stop.

Craigslist’s Missed Connections, perfect for situations like these, has always been a hotbed of unrealized moments that give elusive hope that you may reconnect with someone you shared a few serendipitous seconds with. Yet, it can still be difficult to cull through the vague explanations and ambiguous descriptions. And, who has the time to troll Craigslist every time they pass someone by?

Brooklyn artist and NYU professor Luke DuBois is aiding these lonely hearts through a website that aims to help people rediscover their missed connections in nine different cities (including New York, San Francisco, and London) by using an algorithm that scans listings and estimates potential matches by looking for similarities in language. It scores words according to commonality, focusing more on similarities in specific descriptions. The more specific people’s descriptions, the more likely they will be connected.

The interface itself is fairly simple—a black screen with two listings in piecemeal. Lines manifest between words, searching for similarities between the posts. If “estimated match” exceeds 85%, the program will open up a new tab and users can help play matchmaker. According to Brooklyn Paper, DuBois has connected eight potential couples, but so far no one has written back.

An interesting side effect to this is that it could possibly connect people who were not initially looking for each other, or even connect people who would be a more likely match than you and your subway crush. It’s a kind of automated cupid that slings algorithmic arrows between people with similar vocabulary.

[via Beta Beat]