https://github.com/pavankethavath/music_recommendation_engine

An advanced Music Recommendation System leveraging a Spotify dataset to deliver personalized song suggestions. The project applies KMeans clustering, PCA, t-SNE, and cosine similarity for precise recommendations. Built with a user-friendly Streamlit interface, it showcases data preprocessing, unsupervised learning, and insightful visualizations.
https://github.com/pavankethavath/music_recommendation_engine

clustering elbow-method feature-extraction kmeans-clustering pca recommendation-system spotify streamlit tsne

Last synced: about 1 month ago
JSON representation

An advanced Music Recommendation System leveraging a Spotify dataset to deliver personalized song suggestions. The project applies KMeans clustering, PCA, t-SNE, and cosine similarity for precise recommendations. Built with a user-friendly Streamlit interface, it showcases data preprocessing, unsupervised learning, and insightful visualizations.

• Host: GitHub
• URL: https://github.com/pavankethavath/music_recommendation_engine
• Owner: pavankethavath
• Created: 2024-10-31T18:50:37.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-11-27T16:39:20.000Z (over 1 year ago)
• Last Synced: 2025-03-29T22:25:10.976Z (about 1 year ago)
• Topics: clustering, elbow-method, feature-extraction, kmeans-clustering, pca, recommendation-system, spotify, streamlit, tsne
• Language: Jupyter Notebook
• Homepage:
• Size: 38.6 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# 🎵 **Music Recommendation System**

An advanced **Music Recommendation System** powered by data science and a comprehensive **Spotify dataset**. This project leverages real-world song data, unsupervised learning, and similarity metrics to deliver personalized music recommendations.

![Streamlit Interface](./images/home.png)

---

## 🌟 Overview

This project demonstrates the application of data science to solve real-world problems in the music industry. By preprocessing, clustering, and analyzing song data, the system provides recommendations for users based on the similarity of features like danceability, energy, and tempo. 

---

## 🌟 **Project Highlights**

- **Real Data**: The dataset includes detailed song-level attributes sourced from Spotify, providing insights into audio characteristics.

- **Unsupervised Learning**: Implements **KMeans Clustering** for grouping similar songs.

- **Similarity-Based Recommendations**: Uses **cosine similarity** to recommend songs closest to user preferences.

- **Interactive Interface**: A clean and user-friendly **Streamlit** web application for seamless user interaction.

- **Data Visualization**: Insightful visualizations showcasing clustering, feature correlations, and distributions.

---

## 🚀 **How It Works**

1. **Data Collection**: 

   - The dataset contains song features like `danceability`, `energy`, `tempo`, and more.

   - Metadata such as song names, artists, popularity, and release year are included for richer insights.

2. **Data Processing**:

   - Standardized numerical features to ensure uniformity using **StandardScaler**.

   - Dimensionality reduction using **PCA** for noise reduction and computational efficiency.

3. **Unsupervised Learning**:

   - **KMeans Clustering** groups songs with similar attributes.

   - Optimal clusters are determined using the **Elbow Method**.

     ![PCA Visualization](./images/elbow_method.png)

4. **Recommendation Generation**:

   - **Cosine Similarity** measures the closeness of songs within a cluster.

   - Recommendations are ranked by similarity score.

5. **Interactive User Interface**:

   - Built with **Streamlit** for a dynamic and easy-to-use app.

   - Users input a song name to get recommendations based on dataset attributes.

---

## 📂 Project Components

- **Data Preprocessing and Clustering**:

  - File: `Music_Clustering_and_Recommendation.py`

  - Steps include data cleaning, normalization, clustering, and model saving.

  

- **Recommendation Engine**:

  - File: `recommendation_engine.py`

  - Implements the recommendation logic with a user-friendly Streamlit interface.

- **Saved Artifacts**:

  - `preprocessed_data.csv`: Preprocessed music data ready for recommendations.

  - `kmeans_model.pkl`: Trained KMeans model for clustering.

## 🚀 How to Run the Project

### Prerequisites

Ensure the following are installed:

- Python 3.8 or higher

- Libraries: Streamlit, Pandas, Scikit-learn, Matplotlib, Seaborn, Plotly

### Steps

1. Clone the repository:

   ```bash

   git clone https://github.com/your-username/music-recommendation-system.git

   cd music-recommendation-system

2. Run the Streamlit app:

   ```bash

   streamlit run recommendation_engine.py

3. Open the URL (usually `http://localhost:8501`) in your browser.

4. Enter a song name and get data-driven recommendations!

   

# 📊 Data Features

### **Features in the Spotify Dataset**

- **Audio Features**:

  - `Danceability`, `Energy`, `Acousticness`, `Instrumentalness`, `Loudness`, `Speechiness`, `Liveness`, `Valence`, `Tempo`

- **Popularity**: Reflects the song’s global reach.

- **Metadata**: Duration (in milliseconds), release year, song name, and artist.

### **Preprocessing**:

- Standardized numerical features for uniform scaling.

- Added a `decade` feature for exploratory analysis.

![PCA Visualization](./images/popularity.png)

---

# 🖼️ Visualizations

### **Clustering Visualization (PCA)**:

Clusters are visualized using **Principal Component Analysis (PCA)**, showing clear separations among song groups. Here's an example of the clustering output:

![PCA Visualization](./images/pca.png)

### **Clustering Visualization (t-SNE)**:

To further understand cluster separations, we used **t-SNE** for high-dimensional data visualization:

![t-SNE Visualization](./images/t-sne.png)

### **Streamlit Interface**:

A snapshot of the interactive Streamlit interface where users can input a song name and view recommendations:

![Streamlit Interface](./images/streamlit_app.png)

### **Feature Correlations**:

A **correlation heatmap** highlights the relationships between features, enabling effective feature selection:

![Feature Correlations](./feature_correlation.png)

---

# 🤖 Data Science Techniques

### **Dimensionality Reduction**:

- PCA and t-SNE for reducing noise and improving cluster interpretability.

### **Unsupervised Learning**:

- KMeans for clustering songs based on similar features.

### **Similarity Metrics**:

- Cosine similarity for ranking song recommendations.

### **Model Evaluation**:

- **Silhouette Scores** and **Elbow Method** for validating clustering performance.

---

# 🔍 Key Insights

- **Data-Driven Approach**: Demonstrates the application of unsupervised learning for personalized recommendation systems.

- **Effective Clustering**: Clustering simplifies the recommendation process by narrowing down candidates.

- **Scalability**: Ready for integration with large-scale datasets or additional features.

- **Interactive Visualizations**: Improves transparency and interpretability of the model’s results.

---

# 🎯 Future Enhancements

### **Real-Time Data Integration**:

- Add more metadata or integrate with live Spotify API data in the future.

### **Advanced Models**:

- Explore deep learning techniques such as Autoencoders for feature extraction.

### **Enhanced Clustering**:

- Experiment with algorithms like DBSCAN or Hierarchical Clustering for better performance on sparse datasets.

---

# 💡 Contributions

Data enthusiasts and developers are welcome! Feel free to fork the repository and submit pull requests to enhance the project.

---