Rhythms and Data Science

I have always been interested in both music and technology and the trends associated with them. If you think about 3D TVs or Disco, then you understand that trends come and go.

It is still interesting to study our musical consumption and I intend to do just that throughout this blog using all the tools I learn in my data science bootcamp at General Assembly.

I think music we love and can’t stop listening to is based on feelings but as much as we are different, we are also very similar. Did you know that more than one third of the 30 millions of songs in the Spotify database are in the key of G major, C major, D major and A major ?*

Chart of the keys of all music on Spotify (Kenny Ning, insight.spotify)

But, what is a good song really ? I think it depends on who ask the question and what is the criteria to define the quality of a song. Something I find very interesting and inspiring at the same time is the use of sampling, interpolation and covers in music. Sampling is the act of reusing an artist composition as the basis for a new composition. Take for exemple the most sampled song ever : “Amen, Brother by The Winstons” . It has been sampled 3,118 times*. The original track was released in 1969 or 52 years ago ! James Brown is another example being one of the most sampled artist of all time with 7,413 use of his music in other compositions.

Covers have always been a part of music but sampling is a direct result of advancement in composition methods and thanks to data we can see that some songs live longer than others.

Percentage of samples in popular music (2019 tracklib; state of sampling)

Familiarity and moods as mentioned before are a big part of what make us tick. Another important part is the lyrics. Can we observe a trend of the most popular topics ?

Yes, according to the Atlanta Institute of Music and Media, there are 5 major recurrent topics: Growing Up, Statement of discontent, Friendship, Heartbreak and Death.

What ? Heartbreaks over love songs !

This is the type of insight that I’m looking to share on this blog. Web crawling and visualization are interesting tools that I‘ll utilize to collect a lot more data and to share many more info on the subject I love the most.

In the meantime; I’m going back to Steely Dan !

sources: https://io9.gizmodo.com/a-chart-of-the-most-commonly-used-keys-shows-our-actual-1703086174

https://www.redbull.com/us-en/these-are-the-most-sampled-artists-and-songs-in-music

immersive data science bootcamp @ General Assembly