Swapbeat.com: a streaming, adaptive music service in the browser
Abstract
In recent years, ongoing shifts in the economics of the music industry have driven many artists to reevaluate how they intend to make money, with some artists now deciding that they will give their music away for free if it will increase their popularity and drive more fans to buy tickets to see their live shows. Sites such as SoundCloud.com cater to this dynamic, allowing artists to host their music on the site for unlimited free streaming and download by fans. However, the sheer volume of music being uploaded to SoundCloud makes it impossible for even the most diehard music fans to keep tabs on what has been released and what is becoming popular with other listeners. This paper therefore discusses the creation of swapbeat.com, a browser-based music streaming service. The site operates by first measuring how frequently new songs are being reviewed across 175+ popular music review sites on the Internet. The most frequently reviewed songs are then added to the swapbeat.com site, where score-based playlists of the songs are regenerated at frequent intervals, based on several explicit and implicit measures of sentiment gathered from users as they listen to music on the site. Experiences and observations from the first two months of operation for swapbeat.com are then reviewed, and areas for further work are proposed.