This notebook is an EDA of the "19,000 Spotify Songs" dataset available on Kaggle. It is mostly an excuse to try pandas-profiling on a dataset of a few thousands observations, and carry on the EDA after that.
nb. Pandas-profiling is a library that generates (and export) html EDA reports automatically.
The rendering of Jupyter notebooks doesn't work well with html inside notebooks, and pandas-profil output its reports in html.
Nbviewer, however, works fine:
https://nbviewer.jupyter.org/github/fauconnier-n/EDA-Spotify-19000-songs/blob/master/EDA_19000_songs.ipynb
What it looks like:
