Data-Driven Optimization of Pharma Manufacturing and Distribution

A Streamlit app for pharma supply-chain analytics with live Google Sheets integration, AI-driven chatbot (Google Gemini), demand forecasting (SARIMA & deep learning), visualization, and unit-price prediction.

---

🚀 Features

1. 🤖 PharmaBot Chatbot

   • Ask any question about your shipment data
   • Context-aware RAG pipeline with embedding-based chunk retrieval
   • Structured JSON prompts for accurate, data-driven answers
   • Follow-up memory and “Generate Visualization” toggle

2. 📈 Demand Forecasting

   • Traditional SARIMAX forecasts with reliability scoring
   • Debug info, model metrics (RMSE, MAPE, R²), and confidence indicator

3. 📊 Interactive Visualization

   • Timeline slicers, multi-type charts (line, bar, histogram, pie, box, heatmap)
   • Filter by date, product, vendor, or any dimension
   • Responsive design and download via chart toolbar

4. 💰 Unit Price Prediction

   • RandomForest regression for per-unit price
   • Combined with time-series pack-price forecasting
   • Key driver analysis and prediction intervals

5. 🔄 Live Data Backend

   • Google Sheets API (no manual CSV uploads)
   • Service-account auth for secure, real-time sync
   • Outlier removal built into loader

Website created using Lovable

Sample Website:

URL: https://lovable.dev/projects/7e2a83ed-fd08-480f-a4b6-345796d6fd4e

---

🛠️ Installation & Setup

# 1. Clone the repo
git clone https://github.com/Praj460/Chatbot-for-supply-chain-Analysis.git
cd Chatbot-for-supply-chain-Analysis

# 2. Create & activate a virtual environment
python3 -m venv venv
# macOS/Linux
source venv/bin/activate
# Windows
venv\Scripts\activate

# 3. Install dependencies
pip install -r requirements.txt