Playlist Assist for iTunes

Playlist Assist is a simple aplication that will improve the Party Shuffle feature of Apple iTunes for Windows. Rather than selecting upcoming songs uniformly at random, Playlist Assist identifies and suggests songs in your library that are similar to those already in the playlist.

Playlist Assist makes it easy to build a playlist starting with just a few "seed" tracks. You can use it on both Party Shuffle and regular playlists. As you accept and decline songs, a simple machine learning algorithm identifies what the songs you've accepted have in common, and suggests similar tracks.

The suggestion algorithm considers the metadata stored with each track and also information that iTunes records about your listening habits. It works best if you are a long-time iPod+iTunes user with some playlists already set up.

The current version of Playlist Assist is a sort of awkward, separate add-on program for iTunes. In the future, I'll integrate it more closely with the application; for now, I'm interested in getting feedback on how well the suggestion algorithm works.

Requirements:

· To use Playlist Assist, you must have a recent version of Apple iTunes for Windows. You also need the .NET Framework 2.0 installed; it is likely that you already have it, but if the installation fails for this reason.