Predicting a song's popularity

As part of the course “Data Analysis” from Pontificia Universidad Católica del Perú, a group of students and I applied Machine Learning algorithms, using Python, to uncover which parameters of a song were the best predictors for its popularity.

We analized a data frame whose rows represented Spotify songs, and whose columns contained musical information about the songs, such as tempo, key, time signature, etc, from which we uncovered, using logistic regression, the best song popularity predictors: danceability and loudness.