Beat Box
A novel song recommendatoin system built using Content Based Filtering and Natural Language Processing by Ansh Sachdeva and Shivam Vats
Beat It
Beat Box generates personalized song recommendations based upon users’ playlist. The dataset used in this project comes from a playlist on Spotify.
Demo
Installation
Clone the repository using this command:
git clone https://github.com/shivam360d/Novel_Song_Recommender.git
Data Collection
We performed data collection using Spotipy package in python, where we used a playlist to populate our corpus with around 10000 songs. You can augement the corpus by updating the playlist Id in the jupter notebook named extraction-api present inside the notebooks directory. Please populate the client id and the client secret values in the same notebook. You shall refer this document to get this pair of credentials.
Deployment
Create a virtual environment in python using the foloowing commands:
python3 -m venv venv
source venv/bin/activate (or venv\Scripts\activate if you are using Windows)
Installing dependencies in virtual environment:
pip3 install -r requirements.txt
Running the app:
cd recommendation_app
python wsgi.py
Visit 127.0.0.1:5000 and then beat it!
Click here to learn more about our recommender system.