Data Science of Pop Music: Investigating how the charts have been evolving and what drives them

Pop music is the sound-track to our lives, and it’s one that never stands still. It evolves.  But how do the pop charts change? Why?  And who – if anyone – controls them?  Some say that the artists who make it do; others point to culture-industry (record labels, broadcasters); others simply say that that we – the consumers who actually buy and listen to the stuff – do.

Questions like these are at the heart of this project.  It’s co-supervised by an evolutionary biologist at Imperial College’s Data Science Institute, a computational social scientist at Kings College London, and the head of the BBC’s Research & Development Labs.  In this project, you will use Musical Information Retrieval tools to computationally measure the properties of tens of thousands of pop songs.  Then, you’ll use various time-series and machine-learning techniques to find out how the charts have been evolving and what drives them.  You’ll investigate, quantitatively, how Radio 1 DJs have shaped the charts, and will delve into the vast store of music made by obscure artists across the nation, that never gets heard, but is held by the BBC.  You’ll find out who really makes the music.

You will be based partly at Imperial College’s Data Science Institute and partly at the BBC’s labs; but you will also get training at King’s College Digital Humanities. To do this project, you will need strong data-science skills.  Competence in either R or Python is essential. You could have a business, computer-science, social-science or natural-science background – but you will be expected to grapple with theories and techniques that range from evolutionary biology to sociology and cultural theory – with, perhaps, some physics thrown in.  department. You should also really like pop music!

To apply please contact Prof. Leroi initially.

NOTE: EU students are funded on a fees-only basis