Pandora's podcast recommendation engine launches in beta today

Pandora made its interest in podcasting clear several times this year. Now, the music-streaming company is ready to announce the "Podcast Genome Project," a tool that's both a cataloging system and recommendation algorithm for spoken audio programs. Much like the company's existing Music Genome Project, which categorizes music on hundreds of characteristics to make recommendations to listeners, the Podcast Genome Project (which I'm now abbreviating to PGP). As of today, the PGP is in public beta. For starters, the PGP analyzes and makes recommendations for podcasts based on "over 1,500" attributes, although we don't yet know what sort of characteristics it uses to categorize different shows. Additionally, it also takes user feedback (like thumbs-up or thumbs-down ratings, skips and replays) into account and combines that with machine learning analysis and natural language processing to build a profile for each individual users.