Music Recommender Analyzes Beats and Tweets
Echo Nest is software that recommends related songs in Spotify and other music-discovery services.
It's hard enough sometimes to know exactly what you like about a song. So how can a computer tell? A recent Fortune article provided a peek into the software that powers Spotify's Radio app, iHeartRadio, VEVO's "similar artists" option and other music recommendation services. These are the services that are able to suggest the rapper Drake to a listener, if the service knows its user already likes Kanye West.
The software, called Echo Nest, takes songs in its database and finds those songs' key, tempo and other basic characteristics. Echo Nest then combines that data with context it finds from analyzing tweets and other public online posts about music. It's able to parse comments such as "that song is danceable" and "that singer is Dylanesque," Fortune reported. Echo Nest's database, which its creators got from their clients, includes 34 million songs by 2 million musicians.
The Massachusetts-based startup's list of clients includes MTV, Nokia, Intel and smaller companies, such as Rdio. Pandora, another well-known music recommender, has its own proprietary software and could be an Echo Nest competitor, Fortune reported.
Echo Nest's next moves include continuing to provide its application programming interface for free, for noncommercial uses. The interface, which software developers call an API, allows developers to build interesting apps using Echo Nest data.
The music-matching startup is also working on interesting apps internally. They've looked at matching songs with people's political affiliations and other personal, non-musical tastes.