ML with Spotify
So far, I’ve seen so many Spotify analysis on the net. But i wanted to do something different. So i teamed up with my friends to fulfill our assignment at Sabanci University, Data Analytics master program.
For a solution to come up, there must be a problem. So, what might be the problem with Spotify? Let’s go.
You have liked(i.e pushed the like button) some songs on Spotify but have forgotten to put some of them in your categorical playlists. maybe at first you did it, but later on, just in our case, you didn’t bother with it and it cumulated to tens maybe hundreds that are in your library(liked songs lib) but not in playlist.
And here is another problem. The number of your playlist soared, and some playlists resembled the others, and you are just confused and don’t know where to put a new liked song, actually that’s why the first problem arises, you simply left it un-playlisted.
All you want is, you want to listen to the similar songs altogether.
To sum up: we should run such a model that similar songs should go to the same playlists, so we will build a clustering model for this. and for the new songs, we could just employ the predict method of the model to put them in the relevant cluster, i.e playlist.
We use spotipy api to process playlists. And you could find the necessary files below:
This is a quite a long notebook, so you can carry on the rest on the notebook.
Thanks for reading
Don’t forget to like if you find it useful.