I just finished reading Dataclysm, by Christian Rudder. He's a co-founder of OKCupid, the internet dating site. In the book he basically looks at the aggregate data gleaned from his website as well as other dating sights, facebook, twitter, google and Reddit.

Many of his conclusions are predictable, but now we have the raw data to confirm the conventional wisdom. For example, he cites the chart we've all seen that Men, regardless of age, prefer women ages 20 to 22. Once a woman reaches age 23, she becomes much less desirable for most men.

Another interesting set of data is that the least desirable people are black women. Black women are consistently rated as less attractive than other women, and they receive only 75% of the messages that other women get. Black men also get this treatment, but not as pronounced as black women. (Asian men are also undesirable.)

I think we already knew this, but the interesting thing about the data is the OKCupid user base: it's mostly made up of young, urban, educated users, who describe themselves as "liberals" to a 2 to 1 margin. Thus, this is the very population that was so proud of themselves for voting for Barack Obama, and who are so quick to accuse someone else of being a racist, when their own preferences reveal that they are as race conscious as the people they slander. The only difference is that they lie to themselves about it.

At one point, the author describes an experiment that occurred at OKCupid. They basically shut down profile pictures and had people go on "blind dates." The experiment wasn't successful according to the author. But, while the experiment was running, they noticed that the relative hotness of the participants to each other (or attractiveness disparity) did not affect how much they enjoyed themselves. Thus, if a very hot woman was paired with an unattractive man, and vice versa, they still reported similar levels of "date satisfaction." Women had a good time 75% of the time and men had a good time 85% of the time.

This is a very interesting fact because the data shows us that physical attractiveness is very important when making the initial selection. The more attractive the person, the more messages he or she gets. Thus, people are self-selecting for attractiveness, but the data reveals that it's not really that important; that is, it doesn't affect how much the participants enjoyed themselves. This is counter intuitive, especially here at TRP. The aggregate data shows us that attractiveness, and comparative attractiveness is less important than we think it is.

The author also talks about reddit. He provides a chart that shows the relationship between the 200 most popular subreddits and how isolated the subreddits are. They are arranged by the level of cross-commenting between them. TRP is associated with the following subreddits: TwoXChromosomes, hockey, sex, Hom