Predicting Success for Musical Artists through Network and Quantitative Data

With the rise of Spotify, iTunes, and YouTube, ninety-nine cent songs have largely replaced $20 albums, slashing music sales by nearly fifty percent [1]. Investing in many artists is prohibitively risky today. It behooves music industry executives to leverage available metrics and machine learning techniques to predict whether an artist will be commercially successful in the future. Our goal is to predict whether artists will be successful based on available music industry metadata, namely artists’ importance in the music industry and the public’s response to their music. We collected annual measurements from 2000 to 2012 representing these features and calculated their change over time. We predict the success of an artist in 2013.