ADE 2017: How Hitwizard predicts Spotify success
Amsterdam Dance Event (ADE) always offers a wide range of highly insightful conferences and this year was no different with a particular highlight coming in the shape of a talk from Hitwizard. Designed by Goldmund Wyldebeast & Wunderliebe, Hitwizard is a self-learning computer algorithm that makes predictions on the likelihood of a track appearing in the Spotify Top 200, based on airplay and other unique characteristics.
The AI is currently trained with 2.5 million lines of data and more coming every day. By using a neutral network the AI is able to find predictors for parameters in the data that are very hard to find for humans. Not just the amount of airplay but also time, place, and unique identifiers for a track provided by the Spotify Audiofeatures API gives the AI unprecedented insights in what tracks might become hits.
The talk itself, broadcast from Amsterdam Dance Event 2017 can be found in video form below and featured a highly esteemed panel of music industry personnel, including:
Ard Boer (Goldmund Wyldebeast & Wunderliebe, NL)
Bo Plantinga (Sony Music Entertainment, NL)
Dominique Keegan (Kobalt Music, US)
Jon Davies (Shazam, GB)
Onno van Kemenade (Armada Music, NL)
Eric van der Kwast (TTTManagement, NL)