Spotify Analysis

As an avid music listener and Spotify user, I harbor a personal interest in the data the streaming service makes available. Namely, I wanted to see my own personal listening history with Spotify- using Tableau, I created a brief visualization that summarizes my listening tendencies. I included a filter for ‘Artists’, which shows me the minutes listened vs. the number of tracks I listened to by that artist (one month I could be binge-playing one song for hours on end, and the next I could be exploring more of their work).

An interesting ‘Artist’ filter here is Fleetwood Mac— my favorite band. This past August I spent 263 minutes listening to them; but more interesting than that is the fact that I listened to 110 different Fleetwood tracks. If Spotify didn’t have 12 different renditions of ‘Rhiannon’, how would I have gotten through August? The world may never know.

As an avid music listener, I’m also curious with how trends differ between different areas of the world. After finding a public Spotify dataset via Kaggle.com. After some data visualization using Tableau, we can see not only which countries lead Spotify in streams, but who the top artists are in respective countries, and what their top hits are. It’s no surprise that Drake is Canada’s #1 artist, but he doesn’t even break the Top 10 in Peru— adds perspective!

Finally, some data analysis. Using another Kaggle.com dataset of Spotify songs and difference characteristics (loudness, wordiness, instrumentalness, key, etc.). This short presentation walks us through a few hypothetical questions we all have about music… Is Electronic music or Dance music more ‘danceable’? Does a song’s ‘liveness’ correlate with it’s popularity? And speaking of popularity— is there a ‘perfect recipe’ of characteristics that produce a popular hit? While these questions are explored further in the presentation below, it’s important to close with the point that all art is subjective. There are no calculable secret formulas to what moves or grabs people; all we can do is observe.

Previous
Previous

Machine Learning and House Price Prediction