Posts Tagged similarity
You can usually learn something about a person by looking at what music they listen to. Someone who listens to the Sex Pistols and the Ramones is likely to be from a very different demographic than someone whose favorite artist is Julie Andrews. Of course there are always exceptions to the rule – there are probably a few playlists out there in the world that have both “Anarchy in the UK” and “My Favorite things” but I’m quite sure you won’t be finding a mosh pit at a Julie Andrews concert any time soon.
As we collect more data about what people listen to we begin to learn more about the demographics of listening. Who really listens to Country music? Are they really mostly right-leaning southerners? Are all Hanson fans now 30 years old? To learn how we can answer some of these questions be sure to read Echo Nest founder Brian Whitman’s latest post on Variogr.am about the kinds of predictions we can make about people based upon what they listen to.
This week, The Echo Nest is releasing some new API features that make it easy for developers to build apps that take advantage of this listening data. One new API is Taste Profile Similarity. This API lets you take a seed taste profile (a taste profile is how The Echo Nest represents an individual’s music taste) and find other taste profiles that are similar to that seed. To demonstrate one type of application you can build with this new similarity API, we’ve created a web app called “What’s your stereotype?” This application will look at your music taste (based on your Facebook likes, or your jams from This is my Jam), and tell you which Internet meme best fits your listening style.
Yes, the app will pigeonhole you into a narrow, and probably demeaning demographic. You will probably be offended. Here’s my musical stereotype:
If you want to have your own music tastes pigeonholed like this can try the app yourself at What’s your stereotype? Just remember, you will probably be offended.
To create this app, we identified a whole bunch of Internet memes and personas and made some predictions about the type of music each of these personas would listen to. We then look at the music taste similarity between you and each of the personas – the closest matching one becomes your musical stereotype.
The hardest part about building this app was identifying all of the appropriate Internet memes, predicting the music taste for each meme, collecting images, links and attribution, and most challenging of all, writing the witticisms that accompany each meme. Leading this effort was Matthew Santiago, our chief data quality guy here at The Echo Nest. Matthew organized the meme-dream team to collect and massage all this data. Our highly creative meme-dream team includes Michelle, Nell, Charlie, Alyse, Ryan, Sonja, Nicola, Sam, Roisin, Julie, Sara and Alex.
This app demonstrates what we can do with just a little bit of data about your music tastes. Using the techniques that Brian describes coupled with all the deep data we are gathering around listening habits will help us get a much deeper understanding of your music tastes. This understanding will be key to helping us craft the best music listening experience for you. So, go check out the What’s your stereotype? . I hope you’ll have as much fun with the app as we had in building it.