OkCupid is a dating site.
It organises matches based around self-selecting data. So you’re given a bunch of questions to answer.
What do I think of such-and-such? How important is the answer a potential partner gives to that question to whether I might date them? And so on and so forth.
I know one couple who met and got married via it. No names, no pack drill, but if they want to out themselves in the comments you know who you are.
So far, so lovely.
But as well as matching people by this data, they analyse it. And that’s some pretty funky shit.
OkTrends, which is their irregular blog about Stuff, can give you useful information about things like how the camera that your photo was taken on can improve your attractiveness.
Science. In sexy action.
While the recent publication of When Zombies Attack! Mathematical modelling of an outbreak of Zombie Infection  represents a great step forward in our scientific understanding of the process of a zombie outbreak, their model posesses a flaw which must be addressed before we can truly say we understand the process of infection.
Their model assumes that ζR represents the number of Removed who subsequently become resurrected. Any undergraduate student of Zombie Virology would realise that there is a further state T from which the Removed may become Truly dead, viz. when their brains are destroyed.
As this T state would remove a certain number of individuals from the R state, the transmission rate for the virus would necessarily proceed at a lower rate. This would inevitably lead to a slowed increase in the number of active zombies, thus increasing the length of human survival to a period of time longer than the standard SIZR model would predict.
In addition, in the SIZR-Q model, zombies and the infected would almost certainly be moved to the T category within the length of any serious outbreak as those quarantined are killed permanently by brain-destruction.
Surely this is something that peer-review should have caught.
1: Munz, Hudea, Imad, Smith? 2007