BizInsight AI

BizInsight AI is an enterprise-grade customer feedback analytics platform that automates customer sentiment analysis, groups customer complaints into category-mapped topic clusters, and provides a Retrieval-Augmented Generation (RAG) assistant for querying review data.

This project was recently migrated from a legacy monolithic Streamlit application to a modern microservices architecture consisting of a Next.js 14 frontend and a FastAPI backend.

---

Key Features

• Automated Sentiment Analysis: Instant processing of customer review datasets with metric tracking (Average Sentiment, analyzed review counts, risk thresholds).
• Categorized Complaint Clustering: Unsupervised topic modeling grouping negative reviews into business-relevant categories (Payment, Delivery, Technical, Account, Product Quality, Customer Service, etc.).
• Retrieval-Augmented Generation (RAG) Chatbot: A contextual question-answering assistant that responds to business queries using only the uploaded customer reviews, preventing hallucinations.
• Standalone Guest RAG Sandbox: A public-facing chat interface that does not require user authentication, utilizing a direct ChromaDB integration.
• Structured Alerts: Monitoring of risk levels, negative review spikes, and notifications.

---

Architecture & Tech Stack

• Frontend: Next.js 14, React, Tailwind CSS, Lucide Icons.
• Backend API: FastAPI, Uvicorn, Python, SQLite.
• Machine Learning & NLP:
  • Sentiment Analysis: NLTK VADER.
  • Topic Clustering: BERTopic, HDBSCAN, UMAP.
  • Sentence Embeddings: Sentence-Transformers (using all-mpnet-base-v2).
• Vector DB & RAG Pipeline: LangChain, ChromaDB, OpenRouter (Gemini LLM).

---

Directory Structure

BizInsight-AI/
├── bizinsight_api/          # FastAPI backend application
│   ├── routes/              # API router files (auth, reviews, dashboard, clustering, admin)
│   ├── models/              # Database models and schemas
│   └── main.py              # Backend entry point
├── bizinsight-web/          # Next.js 14 frontend application
│   ├── src/
│   │   ├── app/             # Next.js App Router (dashboard pages, chat sandbox, landing page)
│   │   ├── components/      # UI components
│   │   └── lib/             # API client utilities
│   └── public/              # Static assets
├── rag_api/                 # Core RAG chatbot service logic
├── clustering/              # Clustering algorithms and category mapping
├── database.py              # Database initialization and connection helpers
├── sentiment.py             # NLTK VADER sentiment analyzer wrapper
├── download_model.py        # Utility to download NLTK data and models
├── sync_vectors.py          # Vector store synchronization script
└── requirements.txt         # Backend Python dependencies

---

Installation & Setup

1. Clone the Repository

git clone https://github.com/Prateekiiitg56/BizInsight-AI.git
cd BizInsight-AI

2. Set Up the Backend API

1. Navigate to the project root directory and set up a virtual environment:

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

2. Install the required Python packages:

   pip install -r requirements.txt

3. Download the NLTK and SentenceTransformer models:

   python download_model.py

4. Configure the environment variables by creating a .env file in the root directory:

   OPENROUTER_API_KEY=your_openrouter_api_key_here

5. Start the FastAPI backend server:

   uvicorn bizinsight_api.main:app --host 0.0.0.0 --port 8001 --reload

3. Set Up the Frontend

1. Navigate to the bizinsight-web directory:

   cd bizinsight-web

2. Install Node.js dependencies:

   npm install

3. Start the Next.js development server:

   npm run dev

4. Open http://localhost:3000 in your browser.

---

CSV File Requirements

To upload review datasets, the CSV files must include at least one column labeled review containing the textual customer feedback.

---

Authors & Acknowledgments

• Prateek Singh - AI & Software Developer