“Hello my name is Eric and I am addicted to music.” Needless to say, I was thrilled when I received one of the early invitations to join Spotify (http://Spotify.com) when it launched recently in the US (if you’re not familiar with Spotify, here’s a good introduction). The service offers a catalog of +15,000,000 tracks, and the audio quality has been consistently excellent.
However, there is one area where I find Spotify severely lacking – discovery. Fortunately, I work in the Semantic Web world, and I recently had the opportunity to play around with Seevl.net, a music discovery service that leverages semantic technology. It’s impressive, and I often find myself using Seevl.net to augment Spotify.
So what is Seevl?
Seevl provides new ways to explore the cultural and musical universe of users’ favorite bands and artists, and lets them discover other connected ones, based on a rich set of connections that can exist. The service, free for its users, offers an online discovery user-experience, whether it is by browsing artists, labels or genres, or by combining these features together to find new ones. It also makes its data available to developers that want to build new applications on top of the platform.
As a user, I can go to data.Seevl.net and simply enter an artist’s name. For example, entering “Hugh Masakela” (http://data.seevl.net/entity/cSuAAD6T) serves up biographical information, YouTube videos, and related bands and artists. It is in this last category where Seevl really shines, presenting me with numerous suggestions which at first glance are not necessarily intuitive. For example, I know the name W.C. Handy, one of the related artists presented, but I know Handy only as a songwriter. I had no idea that he also played trumpet and cornet; this I learned from Seevl. There are also numerous artists presented on W.C. Handy’s page of which I had no previous knowledge. What I particularly like here is the way Seevl shows the connections and gives me the tools to explore the topic easily and by following the threads I find most interesting:
By comparison, when I use the iTunes “genius” feature, I’m at the mercy of an algorithm (albeit a powerful one) that someone else built; an algorithm that makes assumptions about WHY I like a particular artist or song (and is not transparent about those assumptions). Seevl, on the other hand, presents me with multiple suggestions, and then tells me why that suggestion is being made. I may like Hugh Masakela because he plays trumpet; or maybe I like jazz; or artists who have recorded on Verve records; or that hail from South Africa… Seevl doesn’t make this assumption for me, but gives me the tools for these serendipitous discoveries.
And that’s just the basic search functionality…
Seevl also offers a “Semantic Search” option, where a user can enter several different facts that she/he wants to search in combination. For example, “Show me all of the artists associated with the instrument ‘harmonica’ and who have their ‘origin’ in Chicago.” Seevl shows me five people. When I change the query to artists who had their “place of death” in Chicago, I get eight results. And when I ask for “artists born in Chicago, who play harmonica, and who have only been performing/recording since 1970,” I get one: Jerry Portnoy.
The company also has a Chrome plug-in for YouTube that displays Seevl information directly on the video page so that you learn about the bands you’re listening to. The plug-in, called “ContexTube,” is available here, and once again, delivered some interesting facts.
First, a typical YouTube page about the “Queen of Soul,” Aretha Franklin:
View on YouTube: http://www.youtube.com/watch?v=wc0bmBRyxK4
And the same page with the plug-in enabled:
Seevl’s default view (above) displays the biographical information of the artist in question, but also provides a small menu allowing the user to explore a “Fact Sheet,” “Topics,” and “Suggested Videos.” This last item (below) returns much richer results than a typical YouTube screen, as it includes biographical information on each artist in the recommended videos.
So Seevl has a lot to offer a music geek like me, but I believe it’s true power and potential for success lies in the partnerships it’s able to strike and the other businesses that leverage what the Seevl team has built.
One final thought: As I sit here with Seevl and Spotify happily running in the background, I can’t help but wish they were integrated. Both have APIs, so maybe there’s a developer out there who wants to take up a challenge. Daniel Ek, Spotify’s founder, may already be on the track of semantic web integration. This paragraph from a recent BusinessWeek interview with Ek (http://www.businessweek.com/magazine/daniel-eks-Spotify-musics-last-best-hope-07142011.html), caught my attention:
“The promise is this,” he continues. “Once you’ve invested in building that library, that’s value.” It’s also, he believes, what people ultimately are prepared to pay for. Spotify has allowed other companies to build tools to import playlists, including from iTunes, and export them as well. To create the universal playlist standard, though, Spotify has to overcome a problem: Any single song by the same artist can have more than one product number. Ek offers the example of The Police’s Roxanne, which can show up on a greatest hits album, a movie soundtrack, a remaster, and the original album, Outlandos D’Amour. Ek wants every song to have a universal resource indicator, or URI, a way for any site or app to call up a unique item on the Internet. “The URI,” he says, “becomes the new MP3. Or the URL. I haven’t made up my mind yet.”
- Automatic Hashtags & Machine Learning: The New Google+
- Gmail, Meet JSON-LD
- Cambridge Semantics Wins SIIA Software CODiE Award
- ADmantX Named "Cool Vendor" in E-Commerce by Gartner