Music has always had a great impact on our mood. This project is an effort made to analyze the same. Predicting the mood of a Person based on the song he or she is listening to and recommending more songs based on the persons mood.
What it does
The user can enter the song either by speaking or by text and our project will predict the mood of the user. The project can predict mood of the person based on the playlist of the person as well. The project can even recommend more songs based on the current mood of the user. Even if the song is mispronounced, the project will search for the right spelling using google.
How we built it
◼ Created a dataset of songs using spotify API and spotipy and requests libraries
◼ Performed exploratory data analysis, feature reduction, feature scaling for preprocessing the data
◼ Performed clustering based on the features of song and labelled clusters by analyzing each cluster
◼ Developed various classifier models to classify the songs based on functional features
Challenges we ran into
Generalizing the model to accept any song and return its features and training our model according to that.
Detecting the song using Speech Recognition was not giving accurate results in the initial stages.
Certain Difficulties faced while using the spotify API
What we learned
Various classification techniques
using spotify API
What's next for Mood Music
Predicting the mood by identifying the song just by the lyrics.
Weekly analysis of the mood of a user by keeping track of the songs the user is listening to.