Music Recommendation System

A Python application which makes predictions on whether a user would like or dislike a song.

Description

Collaborative Filtering is a user-data-based method which seeks the information of every single user's data on the content. Thanks to dataset, the algorithm aims to classify the users by their likes and dislikes. I use Euclidean distance to calculate the most similar other user to the input user. With the data of matched users, I am able to make song predictions.

Music Recommendation System uses a simple model of database of Spotify with 8 users and 24 songs. The reason for this is to learn the collaborative filtering algorithm and make the correct user classification and learn how to work with data in Python.

Libraries

I use Pandas to read and process the data from song-ratings.csv and SciPy to calculate Euclidean distance. Apart from the purpose of program, I use CSV Reader and convert the CSV file to dictionary to check data while I work on calculation.

Dependencies

• Python 3.9.4

• Pandas.py

pip3 install pandas

• SciPy

pip3 install scipy

Authors

Ege Çam
@egecamx

Inspiration & Acknowledgments

• Mosh's Tutorial on ML with Python
• Jack Bandy's article on recommendation systems