Modern streaming platforms are known for their ability to predict the preferences of their users. The music industry in particular poses a complex challenge due to the vastness of different music genres and songs available.
This project aims to model a Bayesian Network from a dataset built by fetching Spotify's API personal data and to experiment with different queries and methods in order to find interesting relationships. Finally, a use case scenario with the final model is presented.
- Dataset.ipynb: notebook that generates the dataset, with all the explanations on how data was processed, and the reasons behind each choice.
- Bayesify.ipynb: main notebook where the models are built, and all the experiments performed.
- spotifyData.csv: the preprocessed dataset used for esimating the CPDs.