Spotify is shaking up the music industry as we know it. Long gone are the days of swapping tapes with friends after school. You won’t be reading about the latest bands in your weekly magazine or tuning into Top of the Pops for the chart countdown. The excitement of music discovery has declined somewhat. Digitisation of music has taken over the industry, and deep learning is playing a huge part in this.
Spotify is the most popular music streaming service out there and as such, harnesses the most valuable asset a company like this can have – data. Over 100 million people use it, with 30 million paid subscribers. 55% also link their accounts to social media. Around 5 million playlists are created or edited daily. In 2015 Spotify users streamed over 20 billion hours of music. Out of the 40 million users of the Discover Weekly playlist, 71% save at least one song each week.
Enter Spotify’s Release Radar. This new playlist features new releases by artists from your individual musical preferences. I don’t know about you, but Discover Weekly is my favourite Monday morning commute pastime.
The streaming giant is capturing the nostalgia that comes with discovering new music by using big data and deep learning algorithms for Release Radar. Spotify has also enlisted a little help from The Echo Nest (2014), a music intelligence agency. Each user has an individual Taste Profile. This is why your recommended playlists aren’t bombarding you with songs based on your occasional guilty pleasure. Instead, the algorithms build clusters around the genres you listen to most.
So what sets Release Radar apart from Discover Weekly?
Discover Weekly uses streaming and playlist data, to tailor a playlist of songs related to your user streaming history. It usually includes some long-forgotten gems you’ve not heard in a while!
But this is where the challenge lies, new releases don’t have streaming or playlist data and they’re largely uncategorized. Spotify is using deep learning techniques to research the audio itself. It pairs with user streaming data to create new music recommendations. Like Discover Weekly, it will evolve and grow the more the user generates data by listening. So far, it’s been incredibly well-received. I’ve found countless new bands to get my teeth stuck into since its launch last week.
Only time will tell if Release Radar can propel Spotify’s success even further. But going off the initial response I’d place my bet on it becoming a weekly essential for many – myself included!
Thanks to Big Data and Deep Learning, whether you’re a belieber, metalhead or even into your Mongolian Throat-Song-Rap (yes, it’s on there!) – Spotify has you covered.
