BlogsLikeThis
The university project I blogged about in my last post has finished. We presented our results last Friday, and were - much to our own surprise - rewarded with the Project Information Retrieval 2006 award. The main reason being that it was a challenge in a completely uncharted area. It’s been a very creative and inspiring project, I must say. I’ve really learnt loads about the blogosphere, programming, and team work.
The process of the past four weeks in a nutshell: we were presented with a data set of 1 million blogs with 10 million posts. The assignment was to automagically find, analyze and visualize communities within this data. This was realized through the development of a fair number of algorithms that analysed link structures and blog content. The final results of our endeavours can be seen at BlogsLikeThis. Most of the website has been implemented by myself (visualization excluded), as well as a part of the underlying algorithms. For the interested ones: a more technical and detailed explanation on the process can be found in our final paper, downloadable on the BlogsLikeThis About page.
One of the great things about blogspace is that word spreads around so quickly. Just a day after the project’s finalization, some posts about BlogsLikeThis started to appear, such as here, here and of course here. I’ve become so enthusiastic about the whole project, that I actually wouldn’t mind spending another four weeks on improving the service (making the data set live, for example, would be a very cool challenge).
