Song Recommendation System based on Mood Detection using Spotify's Web API

When most existing music recommendation systems are content-based, collaborative, hybrid or context-based with a limited songs database, this study examines these systems and designs a new content-based recommendation system, with a vast number of songs. It takes into account the user's current emotion and then recommends songs based on their previous listening history with the help of song features retrieved using Spotify's Web API. The emotion recognition model gives an accuracy of approximately 67% and the recommendation system successfully generates a 20 song playlist based on the user's Spotify listening history and the current emotion detected.