Click here to preview the new Fast Company

Want to try out the new

If you’d like to return to the previous design, click the yellow button on the lower left corner.

How to Data-Mine for Trends Before You Date Around

If you have any interest in your fellow human being, the OKTrends blog run by dating site OKCupid can tell you more about interpersonal relationships than you've learned in decades of being dumped.

In the style of Google Search Trends, the blog mines its user base for data and comes up with astounding results. OKTrends recently analyzed the words inside 1 million member messages, and correlated the content to its rate of response; the point was to try to understand how people are reacting to certain ways they're approached. (In online dating, getting a reply is apparently everything.) Plenty of the dirt the blog uncovers is specific for online daters—netspeak is a turn-off—but a lot of it speaks to larger trends in dating.


For example: OKCupid found that people were deeply uncomfortable with physical compliments in Web messages. Members also preferred writing back to messages with interesting greetings ("howdy") to plain-sounding messages opening with "hi" or "hello." For men, self-effacing humor is a good tact; for all members, bringing up religion at the outset was typically a turn-off. Take note, serial daters.


The site's most recent post is entitled "Your Race Affects Whether People Write You Back," and discusses how the love connection between black men and Asian women seems to be particularly rich. Other posts use OKCupid's extensive survey data, and delve into the prevalence of rape fantasies by state; Nevada is the big winner there. What about the question: Would you murder someone if you were guaranteed not to be caught? Bully for you, Montana.


Check out the rest of the weird science here.

Add New Comment


  • Aly-Khan Satchu

    Well we do all leave an evidence Trail in the c21st and in an Information Century, having more Information is surely a slam dunk proposition. This might well be a Data mining Treasure trove.

    Aly-Khan Satchu
    Twitter alykhansatchu