BookLamp Gets Ambitious With Recommendations

Mac Slocum
March 19, 2008 | Permalink | Comments (0) | Listen

The creators of the BookLamp recommendation system compare their project -- currently in beta -- to the Pandora music discovery service. Like Pandora, source material (book text) is ported into a software program, then analyzed for patterns and sorted into categories. When someone uses the BookLamp recommendation service, the database serves up similar books according to its categorization scheme (more info on BookLamp's system is available in this video).

I ran a BookLamp recommendation query based on George Orwell's 1984, and the service came back with eight suggestions. The most targeted result was the full text of the U.S. Patriot Act (no joke). Other suggestions leaned toward the sci-fi end of the spectrum.

It'll be interesting to see if BookLamp can reach the creators' goal of 1 million books. Sites like GoodReads and LibraryThing use social networking and the power of community to drive recommendations, but BookLamp will need a significant amount of data to reach the same type of critical mass.