Discover Your Next Favorite Audiobook with AI-Powered Recommendations
An intelligent book recommendation system that leverages advanced machine learning algorithms to help users discover their next favorite audiobook from a curated collection of 4,000+ titles.
- Features
- Demo Screenshots
- Technology Stack
- Dataset Overview
- Quick Start
- Usage Guide
- Model Performance
- Project Architecture
- Contributing
- License
- Acknowledgments
- Content-Based Filtering: Uses TF-IDF vectorization and cosine similarity
- Cluster-Based Recommendations: Groups similar books using K-means clustering
- Hybrid Approach: Combines multiple algorithms for optimal results
- 95%+ Success Rate in generating relevant recommendations
- 🔍 Intelligent Search: Find books by title, author, or keywords
- 🏷️ Genre Explorer: Browse 100+ genres and categories
- 👤 Personalized Recommendations: Custom suggestions based on user preferences
- 💎 Hidden Gems Discovery: Uncover highly-rated books with fewer reviews
- Data Insights Dashboard: Explore trends and patterns
- Visual Analytics: Interactive charts and graphs
- Performance Metrics: Real-time recommendation quality tracking
- Export Functionality: Download recommendations as CSV
- Responsive Design: Works seamlessly on desktop and mobile
- Gradient UI: Modern, professional interface with smooth animations
- Horizontal Layouts: Optimized book card displays
- Real-time Updates: Instant recommendation generation
- scikit-learn - Clustering, similarity calculations, and ML algorithms
- NumPy - Numerical computing and array operations
- pandas - Data manipulation and analysis
- NLTK - Natural language processing and text analysis
- Streamlit - Interactive web application framework
- Plotly - Interactive charts and data visualizations
- Seaborn & Matplotlib - Statistical visualizations
- TF-IDF Vectorization - Text feature extraction
- Cosine Similarity - Content similarity calculations
- K-means Clustering - Book grouping and categorization
- Streamlit Cloud - Application hosting and deployment
- 📚 4,002 Unique Books - Diverse collection across all genres
- ✍️ 2,694 Authors - From bestselling to emerging writers
- 🏷️ 100+ Genres - Fiction, non-fiction, self-help, business, and more
- ⭐ 4.46 Average Rating - High-quality, well-reviewed content
- Ratings & Reviews - User ratings and review counts
- Pricing Information - Current pricing data
- Genre Classifications - Multiple genre tags per book
- Detailed Descriptions - Full book summaries and synopses
- Author Information - Complete author details
- Listening Time - Duration for audiobooks
- ✅ Cleaned & Processed - Removed duplicates and handled missing values
- ✅ Standardized Formats - Consistent data types and structures
- ✅ Validated Entries - Quality checks for accuracy
- ✅ Regular Updates - Maintained and refreshed dataset
# 1. Clone the repository
git clone https://github.com/your-username/audible-insights.git
cd audible-insights
# 2. Create virtual environment
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
# 3. Install dependencies
pip install -r requirements.txt
# 4. Run the application
streamlit run streamlit_app.py
# 5. Open browser to http://localhost:8501- **Quick Statistics**: Overview of total books, ratings, and genres
- **Featured Recommendations**: Curated selection of top-rated books
- **Instant Search**: Fast search across titles and authors
---
- Select a Book: Choose from 4,000+ titles using the dropdown
- Choose Method: Pick from Hybrid, Content-Based, or Cluster-Based algorithms
- Get Recommendations: Receive 1–20 personalized suggestions
- Explore Results: View detailed information and similarity scores
- Select Genres: Choose from 100+ available genres
- Set Filters: Adjust minimum rating and result count
- Discover Books: Find books matching your genre preferences
- View Statistics: Explore genre popularity charts
- Set Preferences: Select favorite genres, rating threshold, and price range
- Choose Length: Specify preferred audiobook duration
- Get Personalized Results: Receive custom recommendations with match scores
- Refine Choices: Adjust preferences for better matches
- Rating Distribution: Analyze book rating patterns
- Price vs Rating: Explore pricing and quality relationships
- Author Analytics: Discover most prolific authors
- Cluster Analysis: Understand book groupings and themes
- Success Rate: 95%+ — Reliable recommendation generation
- Average Rating: 4.2+ — High-quality suggested books
- Diversity Score: 0.75+ — Balanced genre and author representation
- Response Time: Sub-second — Fast recommendation processing
| Method | Success Rate | Diversity | Novelty | Quality | Use Case |
|---|---|---|---|---|---|
| Hybrid | 95% | 0.75 | 0.65 | 4.3 | Best Overall |
| Content-Based | 92% | 0.68 | 0.58 | 4.2 | Similar Books |
| Cluster-Based | 88% | 0.82 | 0.72 | 4.1 | Genre Discovery |
- Precision@10: 0.78 — Relevant recommendations in top 10
- Recall@10: 0.65 — Coverage of relevant items
- Diversity Score: 0.75 — Variety across genres and authors
- Novelty Score: 0.65 — Discovery of lesser-known gems
- Serendipity Score: 0.42 — Unexpected but relevant finds
audible-insights/ ├── 📱 streamlit_app.py # Main Streamlit application ├── 🤖 recommendation_classes.py # ML recommendation algorithms ├── 🛠️ app_utils.py # Utility functions ├── 📋 requirements.txt # Python dependencies ├── 📖 README.md # Project documentation ├── 🚫 .gitignore # Git ignore rules ├── 📊 data/ # Dataset files │ └── streamlit_dataset.csv ├── 🧠 models/ # Trained models │ ├── combined_tfidf_matrix.npy │ ├── content_recommender.pkl │ ├── cluster_recommender.pkl │ └── hybrid_recommender.pkl └── 📓 notebooks/ # Development notebooks ├── 01_data_preparation.ipynb ├── 02_data_cleaning.ipynb ├── 03_exploratory_data_analysis.ipynb ├── 04_nlp_and_clustering.ipynb ├── 05_recommendation_systems.ipynb ├── 06_model_evaluation.ipynb ├── 07_streamlit_app.ipynb └── 08_summary_and_conclusion.ipynb
- Data Collection: Gathered Audible audiobook metadata
- Data Cleaning: Processed and standardized 4,000+ book records
- Exploratory Analysis: Analyzed patterns, ratings, and genres
- Model Development: Built and trained recommendation algorithms
- Evaluation: Tested models with multiple performance metrics
- Deployment: Created production-ready web application
- Feature Engineering: TF-IDF vectorization of book descriptions
- Similarity Calculation: Cosine similarity for content-based filtering
- Clustering: K-Means algorithm for grouping similar books
- Hybrid Method: Weighted combination of multiple approaches
- Evaluation: Comprehensive testing with precision, recall, and diversity metrics
This project is licensed under the MIT License – see the LICENSE file for details.
- Audible: Audiobook metadata and catalog information
- Open Source Community: Libraries and frameworks used
- Netflix: Recommendation system architecture
- Amazon: E-commerce recommendation patterns
- Spotify: Music discovery algorithms
- Streamlit: Web application framework
- scikit-learn: Machine learning and clustering
- Plotly: Interactive data visualization
- Data Science Community: For tutorials and best practices
- Open Source Contributors: For maintaining excellent libraries
- Beta Testers: For feedback and improvement suggestions