Back to the past to charter the vinyl electronic market: A data mining approach

Abstract

This study focuses on perhaps the most iconic media format of all time, the vinyl record. By adopting a data mining approach, the goal is to understand which factors involved in the buying and selling of vinyl, influenced its price, with the initial hypothesis considering record labels and popular rankings to be some of the most contributing variables. To be able to evaluate it, four datasets were created in an endeavor to represent recent and past records of two different genres, Rock and Jazz, by extracting data from Discogs’ marketplace and Billboard’s Hot 100 chart. Such approach unveiled that an artist’s presence in the charts and their labels belonging to one of the ‘Big three’ do not always dictate their records at highest prices. The results also showed that features which measure popularity become more relevant in the ‘era’ where the record’s genre is more popular and that big record labels have been losing market share to an increasing number of independent labels.