I wrote a little crawler that scraps pretty much all of the data for each entry of ld26 (compo and jam). I provide a simple search-interface and an API if you want to do data analytics. The search interface let’s you do simple keyword search in the title and description, filter by user, filter by link label (e.g. web, desktop, etc.), and other fancy things. I’ll provide info on the API tomorrow, to tired…
I wrote this for various reasons. For one, i wanted to know how many entries are using libgdx. I also wanted to be able to track comment count history for individual entries. My gut feeling is that one can predict the winner of LD48 within the first few days of voting merely based on the comment count. Finally, i have a little research project going on where i’ll try to rank LD entries based on various characteristics, then evaluate the ranking with the actual voting results.
In any case, maybe this is useful to someone else as well. Source is available here (JSoup, Jersey, JQuery, Bootstrap).
