Rhythmify is a music analytics Flask web application that leverages the Spotify API for generating playlists based on recommendations and helps us in creating a user analysis of their music tastes. This web application is deployed on Heroku.

Check out my medium article as well

• Features of the application
• Demo of the application
• Application Stack
• How to run the application?
• Code Improvements
• References

• Content Based Recommendation system using Tfidf Vectorizer and cosine similarity
• User Dashboard using Plotly
• Playlist Generator using Spotify API

1. Authorization into Spotify Account: For this, it is required to create an account in Spotify Web API and get a Client_ID and Client_Secret_Key, additionally update the redirect URI including the port in the Spotify API user dashboard. For more details on the authorization procedure, please have a look at : https://developer.spotify.com/documentation/web-api/quick-start/ For more details on the API reference end-points please have a look at: https://developer.spotify.com/documentation/web-api/reference/

2. Playlist Generator: I had incorporated Spotify-API and used the playlist API endpoints to extract my playlist data. Once the playlist is displayed, a new playlist can be generated based on the few tracks selected from the previously saved playlists. I had used the recommendation function provided by the Spotify API to generate the new playlist.

3. User Dashboard. This is a music analytics dashboard created using Plotly. I have analyzed my top tracks and artists and created various visualzations.

4. Recommendation System. A content-based recommendation system was created to recommend similar artists and songs. For recommending artists, I had leveraged cosine similarity based on genre and tracks and Tf-idf vectorizer to predict the artists based on similarity scores. I had used the Spotify Kaggle dataset Similarily, for songs I had used the same technique, but this time I created a tfidf matrix based on the artists.

• Backend: Flask, Jinja2, Python
• API used: Spotify-API
• Frontend: HTML, CSS, Javascript
• Libraries: Numpy, Pandas, Scikit-Learn
• Visualization: Plotly
• Deployment: Heroku

• Clone the repo into your local machine

git clone https://github.com/Srijha09/Rhythmify.git

• Create a Developers account and set up a new project in Spotify API Dashboard https://developer.spotify.com/dashboard/login
• Install requirements

$ pip install -r requirements.txt

• Once the application is cloned and the account is setup, run the flask app

$ export FLASK_APP="app_name"
$ export FLASK_ENV=development
$ flask run

• Expand the Spotify Dataset for more songs
• Create Flask Blueprints
• Lyric2Vec and Audio Modelling Recommendation system
• Mood Based Song Prediction

1. https://developer.spotify.com/documentation/web-api/quick-start/
2. https://github.com/Deffro/statify
3. https://github.com/AcrobaticPanicc/Spotify-Playlist-Generator1