Repository containing the code of the Analysis of Popular Music project

The article of the project can be found on my website.

Whether one likes it or not, music is an art that does not let people indifferent. At a time when listening to music has never been easier, I felt interesting to have a closer look at what are the characteristics of popular music and popular artists over the last half-century.

To do that, I am using data from the Billboard Hot 100 year-end raking scrapped from Wikipedia. All the data analysis phase has been done in an iPython Notebook using Pandas and NumPy for the data wrangling steps. Scikit-learn has also been used for some machine learning tasks.

Several interactive visualizations have been made in the article. They are all using d3.js.

To have access to the tracks acoustic attributes, the artists details and the tracks 30s preview, I have chosen to use different APIs:

• The Echo Nest API (for the acoustic attributes and track details)
• LastFM API (for the artist details)
• Spotify API (for the track preview)