The founders of streaming music site and app have warned that tracking the tracks users dislike is as important for recommendations as understanding what they do like. Musicovery has evolved from Liveplasma, the now-defunct music recommendation site that was founded in 2005. Whereas Liveplasma suggested similar artists based on Amazon recommendations, Musicovery suggests tracks based on mood (Calm/ Energetic, Dark / Positive), with each track listened-to by an expert to give it a position on the axis. Users can further influence the position by rating tracks as “Like” or “Don’t Like”. When Musicovery mapped the songs that people like the most, the results represent pop hits from the 90s and 00s – Wonderwall by Oasis, Hey Ya by Outkast, Feel Good Inc by Gorillaz. But when they analysed the relationship between Like and Don’t Like, to find out the songs that had a very low ratio of Don’t Like, the recommendations are totally different. Ain’t No Sunshine by Bill Withers, Stand by Me by Ben E King, Blue Moon by Julie London – classic songs with universal appeal. “It raises issues regarding the way search engines or recommendation engines work”, argue Castaignet and Vavrille.
Tools: platforms to help you reach new audiences
Tools: Kaiber
In the year or so since its launch, AI startup Kaiber has been making waves,…