Overview
This AI-powered chatbot enhances customer experience for a retail store by handling product inquiries, order tracking, recommendations, and FAQs. Built with spaCy for natural language processing (NLP) and FastAPI for a robust backend, it understands diverse user queries and responds intelligently. The React.js frontend delivers a sleek, interactive UI, making it a seamless tool for shoppers.
Data Sources
- SQLite Database: Stores product details (e.g., iPhone 15, iPad Pro) and order tracking data (e.g., order IDs, statuses)
- CSV Files: Sample product and order data for testing (see
csv database folder)
Technologies Used
- Frontend: React.js (UI), CSS (styling), Node.js (dependencies)
- Backend: Python, FastAPI (API), spaCy (NLP), Scikit-learn (recommendations), Regex (input parsing)
- Database: SQLite (lightweight storage)
Key Features
- Product Inquiry: Check availability, prices, and details (e.g., “What’s the price of iPhone 15?”)
- Order Tracking: Real-time order status by ID (e.g., “Track my order #12345678”)
- Recommendations: Suggest similar products (e.g., “Recommend a phone like iPhone 15”)
- FAQs: Answer store policies and hours (e.g., “What’s your return policy?”)
- NLP: Understands varied phrasing (e.g., “Is MacBook Air available?” vs. “Do you have MacBook Air?”)
Natural Language Processing
- Intent Recognition: spaCy identifies user intent (e.g., inquiry, tracking) from text
- Entity Extraction: Regex and spaCy parse specifics (e.g., “iPhone 15,” “#12345678”)
- Response Generation: Matches intents to predefined or dynamic responses
- Recommendation Engine: Scikit-learn suggests products based on similarity
Performance
- Accuracy: High intent recognition with spaCy’s pretrained models
- Response Time: FastAPI ensures sub-second responses
- Scalability: SQLite suits small-scale retail; easily upgradable to PostgreSQL
Challenges Overcome
- Parsed diverse user inputs with Regex and spaCy for robust NLP
- Balanced lightweight SQLite with functional order/product tracking
- Integrated frontend (React) and backend (FastAPI) for seamless UX
Benefits
- Customer Convenience: 24/7 self-service for inquiries and tracking
- Engagement: Personalized recommendations boost sales
- Efficiency: Reduces staff workload on routine queries
Sample Interactions
**Product Inquiry:**
User: “What’s the price of iPhone 15?”
Bot: “The iPhone 15 is priced at $799. Would you like more details?”
**Order Tracking:**
User: “Track my order #12345678”
Bot: “Order #12345678 is shipped and arriving tomorrow!”
**Recommendation:**
User: “Suggest a good laptop”
Bot: “I recommend the MacBook Air—lightweight and powerful. Similar options: Dell XPS 13.”
**FAQ:**
User: “What are your store hours?”
Bot: “We’re open 9 AM - 9 PM, Monday to Saturday!”
Future Improvements
- Add voice input with speech-to-text (e.g., Web Speech API)
- Scale database to PostgreSQL for larger inventories
- Enhance recommendations with collaborative filtering
Demo
Try It
Clone the repo, set up the backend and frontend, and chat away! Check GitHub for instructions.
Visit GitHub Repo