Smitten by Cupid (or, more accurately, by Cupid's data...)

I love it when lots of people put lots of personal information on the Internet so that social scientists and researchers can luxuriate in all that data. Ever since I joined Facebook four years ago, I've been intrigued by the massive data caches that social networking sites create. At least that's how I justify the unfortunate amount of time that I've spent on Facebook in those last four years.

I just spent about 1.5 hours (that I should have spent on homework... oh well) reading through this fascinating blog. In it, the creators of the OKCupid dating site do some great statistical analysis on the large population of users on their site. It's really interesting to read through all the variables that they consider, the types of statistical methods that they use, and the (often surprising) results that they discover. And the user/reader comments on the posts are quite revealing as well.

My favorite post was called "Flowchart to My Heart", in which the author maps out the considerations which would lead him to a decision on going out/not going out with someone. Of course, individual flow charts will vary widely, depending on which questions are included, how many answers to each question are possible, and how the questions are ranked/positioned in the chart. It looks like a fun little exercise- I may try it myself. A little recreational reductionism never hurt anyone, right?

The most recent post is called "The REAL Stuff White People Like." It gives ranked lists of the most common words and phrases that people put in their profiles, based on racial and gender categories. I'm definitely more of a qualitative-methods girl than this, usually; really, who thinks that complex phenomena like "race" and "gender" can be as neatly parsed out as they are for these statistical purposes? I have to say it, though: reading through this blog made me want to sign up for OKCupid simply so that I could be a little pixel in some of these ultra-cool graphs.