Spotify Recommender System

This project is a Spotify Recommender System that generates song recommendations based on a user's playlist. The system uses audio features provided by the Spotify API to create a feature vector for each song and employs cosine similarity to find songs that are most similar to the ones in the user's playlist.

Features

• Load and process data from Spotify.
• Generate feature vectors for songs in a playlist.
• Calculate cosine similarity between songs.
• Recommend top 40 songs not in the playlist.
• Visualize the cover art of the recommended songs.

Prerequisites

• Python 3.8 or later
• Spotipy
• Pandas
• Scikit-learn
• Matplotlib
• Scikit-image
• streamlit
• Jupyter Notebook (for running the notebook interactively)

Dataset

https://www.kaggle.com/datasets/yamaerenay/spotify-dataset-19212020-600k-tracks/data

Installation

1. Clone the repository:

   git clone https://github.com/snigdhasv/spotify-playlist-recommender.git
   cd spotify-recommender

2. Install the required packages:

   pip install -r requirements.txt

Usage

Spotify API Setup

1. Create a Spotify Developer account and register your application to get your client ID and client secret.

2. Set up your environment variables for Spotify API credentials:

   export SPOTIPY_CLIENT_ID='your-spotify-client-id'
   export SPOTIPY_CLIENT_SECRET='your-spotify-client-secret'
   export SPOTIPY_REDIRECT_URI='your-redirect-uri'

3. If you're running the app then enter client id and client secret directly on the app.

Running the Notebook

1. Open the Jupyter Notebook:

   jupyter-lab

2. Run the cells in the Recommender.ipynb notebook to load data, generate recommendations, and visualize the results. (add client id and secret key for spotify api)

Running Streamlit website

1. Go to cloned directory in command prompt and run the following command:

streamlit run app.py