An implementation of a Spotify-like song radio based on Word2Vec. Based on a large set of crawled playlists and using those playlists as sentence equivalents. The Word2Vec algorithm then produces a vector per song.
Try it out below:
The Word2Vec approach over crawled playlists results in assigning a 100 dimensional vector to each song in the database. Currently there are only 300 000 songs - finding enough playlists if you are not Spotify is hard. When one song is entered, we just find the songs closest to that vector. If there are two songs entered, we find the mid-point and use that to find closest neighbors.