Seevl: Using Linked Data to Reinvent Music Discovery

Seevl is a new website which uses semantic web technology to provide intelligent musical recommendations and to explore the cultural and contextual landscape of music. By exploring the different connecting factors between various artists, be they based on stylistic similarities, shared personnel, or a chronological coincidence, Seevl, (which is in beta at the moment,) can provide the user with recommendations based on these factors, which the user can then follow or ignore, based on their own preferences.

Dr. Alexandre Passant is the CEO of Seevl and a research fellow at the Digital Enterprise Research Institute at NUI, Galway. Alexandre says that idea for Seevl arose out of not only a love of music, but also a desire to know the contributing influences and factors behind that music.

“I think that there is no ideal or complete music recommendation system, because music is something very personal, and there are lots of reasons why you like a band. Two people may like the same band, but may not like the same suggestion, because they like the band for different reasons.

“I collect lots of music, buying records and books about music. When I buy music or read books about music I want to get more than the music itself and get every fact about the artist; the influences, the cultural universe, why they came to that particular genre and so on.”

One of the often-lamented by-products of the digitisation of music is the demise, or at least the marginalisation, of the sleeve note. People must often perform their own independent research if they wish to find out what studio an album was recorded in, or who performed the guitar solo on a certain track, for example. There are a number of avenues through which one might find out this information, but Alexandre believes the variety of sources available can actually prove to be an obstacle for some.

“When I want to find information today about a band, I can go to Wikipedia, I can go to Last FM, I can go to Freebase, but that’s the problem; I can go to several sites. We want to provide a website where you can find all the information, so that’s why we started to aggregate information from all these websites and build artist profiles using all this data.”

The way in which Seevl recommends music is different from, for example, Apple’s iTunes Genius, explains Alexandre, focusing on the characteristics of the music through meta tags, and not whether others have liked it.

“These websites will recommend things that other people buy or listen to when they buy or listen to that thing. So they will say that 90% of people that are listening to Snow Patrol, are listening to that, so you should listen to that. We are doing something completely different, we are getting all this information about the band, so as I said, the genre, the members, the influences and so on, and based on that we suggest other bands that share the same characteristics.

“We are not listening to the music, we’re just getting meta-data. So if someone says that “this band is a punk rock band signed to a label from say, New York or San Francisco, we may suggest to you other punk rock bands from San Francisco, because that might be the reason why you like it.”

In allowing the user to see the reasoning behind the site’s recommendation, Seevl takes into account the personal relationship many people have with music, and does not presume that all people follow the same thought processes in their music choice.

“We have the semantic search feature as well, where you can combine different features. So you can say “I want to find all Irish bands playing punk rock between 1995 and 1997 that have been on this label”, and then you get the results.”

The team behind Seevl have made their data reusable, in the hope that developers will develop applications that incorporate this semantic technology.

“Our focus is to get a really nice knowledge page of music with these recommendations and these facts, but then we want people to, for example, create an iTunes player that will just play recommendations, that kind of thing.”

Potential developers can visit Seevl’s Developer Zone, register for a developer key, and the source code and access to the API is free for non-commercial use.

At the moment, the information used on Seevl is collated from sites such as Wikipedia and the New York Times, but Alexandre hopes to allow user-created content in the future. This will not only expand the knowledge-base of the site, but will also create opportunities for  bands that fall beneath the radar of these other sites.

“On Wikipedia, if you’re a small band, you will be rejected. Wikipedia have this policy that you have to be known well enough to be on the site. If we open the data to let users create content, then bands like that would be able to put their information on the website.

“We are all mostly listening to alternative music and small bands and so on, so the more of these bands that we can have there and help to get promoted and get known, the better.”

For now, the Seevl team (The name, incidentally, “doesn’t mean anything, but if you read it backwards it says Elvis!”) are working hard to get the site out of beta, and user feedback plays an important role in this.

We want to do things that people want, so we want to get feedback from users to see what they want, we don’t want to invest time in building a feature that nobody will use. We ask our users, that’s why we’ve got a GetSatisfaction page, so we can get feedback. There are lots of things that we are gathering feedback on before implementing.”

Alexandre will be presenting Seevl at a workshop at the 130th annual Audio Engineering Society Convention in London on Friday, May 13th.

4 thoughts on “Seevl: Using Linked Data to Reinvent Music Discovery

  1. “Seevl … the name, incidentally, “doesn’t mean anything, but if you read it backwards it says Elvis!” ” :: Yay!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s