I stumbled on an interesting website called Personas; it’s part of the Metropath(ologies) exhibit, that’s currently on display at the MIT Museum by the Sociable Media Group from the MIT Media Lab. It uses natural language processing to create a data portrait of your online identity.
I like the idea, but it didn’t really give any good results. I tried it out a couple of times, and each time I got different results. I understand that I’d get different results for “Melinda Seckington” vs “Miss Geeky”, but I got several different results for “Melinda Seckington”, which just doesn’t make sense. I also don’t understand which websites it analyzes to get its data from. Here’s my results:
Little bit weird, right? I have no idea where terms like medicine, illegal, religion, committees come from; I don’t think I’ve ever written about stuff like that. And even things like politics, and music, I practically never write about that!
The people behind it know that it doesn’t give good results though, and see it as a way to make people think about online identities:
In a world where fortunes are sought through data-mining vast information repositories, the computer is our indispensable but far from infallible assistant. Personas demonstrates the computer’s uncanny insights and its inadvertent errors, such as the mischaracterizations caused by the inability to separate data from multiple owners of the same name. It is meant for the viewer to reflect on our current and future world, where digital histories are as important if not more important than oral histories, and computational methods of condensing our digital traces are opaque and socially ignorant.
It is an interesting idea, and I understand the concept behind it. Still, I’m wondering if there is a site out there that truly analyzes our online identities. I’d love to see something that takes tags and favourites from my Delicious, YouTube, Flickr, etc, adds an analyzation of my blog, Twitter and whatever other websites I’m on, to finally create a nice visualization of my online presence. Does something like that exist already?