song_recommendation (Music Recommendation System)

File Structure

The project has been refactored for improved modularity and maintainability:

• main.py: Main script that orchestrates the entire application
• data_processing.py: Handles data loading and initial processing
• recommendation.py: Core recommendation engine
• chatbot_interface.py: Manages the Gradio-based chatbot interface

Features

• Personalized song recommendations based on user input
• Interactive chatbot interface for music discovery
• Integration with Spotify API for up-to-date song information
• Natural language processing for understanding user queries
• Comprehensive music recommendation system based on song similarities

Installation

1. Clone the repository:

   git clone https://github.com/your-username/fizy_muzik_onerme.git
   cd fizy_muzik_onerme
   

2. Create and activate a virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`
   

3. Install required dependencies:

   pip install -r requirements.txt
   

4. Set up the .env file:
   • Create a file named .env in the root directory of the project
   • Obtain Spotify API credentials: a. Go to https://developer.spotify.com/ and log in or create an account b. Create a new application to get your client_id and client_secret
   • Add your Spotify API credentials to the .env file in the following format:

     SPOTIFY_CLIENT_ID=your_client_id_here
     SPOTIFY_CLIENT_SECRET=your_client_secret_here
     

   • Keep your .env file secure and never share it publicly
   • Note: The .env file is crucial for the project to interact with the Spotify API

How to Use

1. Obtain Spotify API credentials:

   • Create a Spotify Developer account at https://developer.spotify.com/
   • Create a new application to get your client_id and client_secret
   • Add these credentials to your .env file as described in the installation steps

2. Prepare your data:

   • Place the Spotify dataset CSV files in the ./data/ directory
   • Ensure you have data.csv, data_by_genres.csv, and data_by_year.csv

3. Run the main script:

   python main.py
   

4. The script will perform the following actions:

   • Load and process the data
   • Launch the Gradio interface

5. Open the provided Gradio interface URL in your web browser

6. Interact with the chatbot:

   • Enter song names (optionally with years) for recommendations
   • Ask general music-related questions
   • Explore recommended songs and their features

Dependencies

• pandas, numpy: Data manipulation and analysis
• scikit-learn: Machine learning algorithms for preprocessing
• spotipy: Spotify API integration
• gradio: Web interface for the chatbot
• transformers, torch: Natural language processing for the chatbot

Data Sources

The project uses the Spotify dataset available on Kaggle. You can find more information about the original dataset here.

Development

This project is continuously evolving. Future plans include:

• Enhanced chatbot integration for a more intuitive user experience
• Improved recommendation algorithms incorporating user feedback
• Extended genre and mood-based recommendations
• Integration with more music streaming platforms

Visualization for Researchers

While the main functionality of song_recommendation repo focuses on the chatbot interface and music recommendations, the project originally included various data visualization features that may be of interest to researchers in the field of music information retrieval and recommendation systems.

These visualization features, although not part of the main chatbot functionality, can provide valuable insights into music trends, genre characteristics, and the underlying structure of the dataset. They include:

1. Distribution of songs across decades
2. Evolution of sound features over time
3. Top genres based on popularity
4. Feature correlation with song popularity
5. Clustering of genres and songs

Researchers interested in exploring these visualizations can refer to the visualization.py and clustering.py files in the project repository. These modules contain functions for creating various plots and performing clustering analyses.

To use these visualization features:

1. Ensure you have the necessary additional dependencies installed (matplotlib, seaborn, plotly, yellowbrick)
2. Import the required functions from visualization.py and clustering.py
3. Call these functions with the processed dataset to generate visualizations

Please note that these visualization features are optional and separate from the main recommendation system. They are provided for research purposes and to offer a deeper understanding of the dataset and the factors influencing music popularity and recommendations.

Contributing

Contributions to improve fizy_muzik_onerme are welcome. Please feel free to submit pull requests or open issues to discuss potential enhancements.

License

[MIT]

Acknowledgements

• Original Kaggle project by Vatsal Mavani
• Spotify API for providing music data
• Gradio for the web interface framework
• The open-source community for various libraries and tools used in this project