A few years ago I did a lot of reading about algorithms and machine learning as it related to the arts and popular culture. What immediately sprang to mind when I read the opening of “Quantitative Formalism: An Experiment” was an article from WIRED in 2011 about at team from The University of Bristol that worked on developing an equation that could predict a hit song.
http://www.wired.com/2011/12/hit-potential-equation/
At the top of the article you’ll see a video that shows the “evolution of musical features” as they relate to hit songs. Since we will soon be considering ways to display results from our data sets I thought this might be interesting to take a look at.
The short article considers both the Bristol team’s work and other similar projects related to predicting the popularity of new pop music. While this is not scholarly work, I thought it was interesting to share and consider how this type of enquiry is being used outside of the academy.
This project reminds me of the the NYTimes video that mapped the making of that recent Justin Bieber/Skrillex hit: http://www.nytimes.com/interactive/2015/08/25/arts/music/justin-bieber-diplo-skrillex-make-a-hit-song.html?_r=0. It went viral.
Oh, right! It’s such an interesting way to visualize the song elements.