We can use data preprocessing recipes to implement dimensionality reduction and understand how features of songs are related to their performance on the Billboard Top 100 chart. Check out the code on my blog: https://juliasilge.com/blog/billboard-100