How vinyl records & big data make Spotify sound better

The next time you’re listening to Spotify, you might want to thank the world’s geeky record collectors and their shelves full of dusty vinyl albums. That’s because some of Spotify’s knowledge about music is powered by data that has been accumulated by DJs and other fans of the analog music format, thanks to a new deal between the big music data provider The Echo Nest and online music resource

The deal, which was announced Tuesday morning, gives The Echo Nest access to discography information about more than two million artists and 430,000 record labels. Many of these hail from the world of electronic music, but Discogs also lists rare and unreleased singles from stars like Kanye West and obscure remixes of classics from bands like Depeche Mode.

Discogs users have in many cases been using the site to catalog their entire collections, or to put part of those collections up for sale. And since every little difference matters to a collector, you’ll often find dozens of different versions of a single album. It basically doesn’t get more detailed than this. “Discogs is an amazing resource that has been around for a long time,” The Echo Nest’s co-founder and CTO Brian Whitman told me during a phone call this week.

The Echo Nest uses this data to improve listening recommendations it delivers to Spotify, iHeartradio, MOG, Vevo and others. The company has been aggregating data from music blogs for around seven years now, and later added Facebook(s fb), Twitter and other sources to the mix. The Echo Nest has more than five billion unique data points, Whitman told me, and has essentially become the big data resource of record for the digital music space. Around 340 applications use Echo Nest data to help their customers find new music and learn more about the music they’re listening to.

Spotify, for example, uses data from The Echo Nest to power its Spotify radio, a feature the music service launched in March to compete with non-interactive offerings like Pandora. Speaking of Pandora: The service has a very different take on music recommendation, relying heavily on experts to classify tunes based on its own music DNA taxonomy. “It’s great what they do, but they can’t scale,” argued Whitman, adding that Pandora only offers around a million tracks. The Echo Nest on the other hand has information about more than 30 million tunes.

So what’s next for The Echo Nest? Whitman told me that the company’s next big challenge is social. Using Facebook and Twitter to gauge which records people are talking about is one thing; utilizing all this data to deliver personalized music recommendations is a whole different challenge.

Of course, once you have figured out to make sense of all that data, you could just as well use it to recommend movies, books or video games – but Whitman said his team of 40 wasn’t interested in expanding its focus at all. “We are a music company, we are never gonna make movies,” he said.

The Echo Nest is based in Somerville, MA and has raised a total of $8.3 million in funding.

Image courtesy of Flickr user BGSU University Libraries.