Semantic Recommender
Built a semantic book recommender using emotion analysis, vector embeddings, and an interactive Gradio dashboard — inspired by a JetBrains tutorial on machine learning workflows
SmartLit: Book Discovery Dashboard
Book Recommender is an intelligent app that helps readers discover new books tailored to their tastes by analyzing their reading history and preferences.
This project leverages Python (pandas, scikit-learn) for building the recommendation algorithms and Flask for serving the web interface. The recommender system evaluates user ratings and book metadata to combine collaborative and content-based filtering results. Data visualization dashboards (matplotlib, seaborn) highlight top books, genres, and user patterns.project demonstrates practical skills in EDA, modeling, and deploying ML-powered web apps.
The primary challenge was ensuring relevant book suggestions for users with limited or very diverse reading histories. Developing an accurate recommender system required handling sparse rating data, identifying meaningful user preferences, and filtering out less relevant recommendations.
The system uses both collaborative and content-based filtering to give each user tailored book suggestions. Built with Python and Flask, it adapts recommendations even with limited data and keeps the experience interactive.
Key Features
- Real-time Data Visualization
- Zero-Shot Text Classification
- Semantic Similarity Search
- Emotion and Sentiment Analysis
- Interactive Gradio Dashboard
- Scalable and Extensible Pipeline