Great Lakes PGP Data Science & GenAI | Batch Sep 2025–Apr 2026
An AI-powered decision support system for India Post officers to analyse district postal infrastructure, understand delivery performance, and get targeted recommendations.
Deployed on Streamlit Cloud → https://postbot-5gnefgmmr9pq4p36fmd4jz.streamlit.app/
| Feature | Description |
|---|---|
| 📦 Real Data | Actual BO / PO / HO counts for 754 districts across 36 states |
| 📊 Delivery Analysis | Current delivery rate + performance tier (Low / Moderate / Good / High) |
| 💡 Smart Suggestions | Exact number of Branch Offices to add to reach next tier |
| 🤖 AI Chatbot | Ask anything — powered by Groq Llama 3.3 70B (free) |
| 🔐 Secure Login | Officer ID + password authentication |
PostBot/
├── app.py ← Main Streamlit application
├── auth.py ← Officer login & authentication
├── chatbot.py ← Groq AI chatbot brain
├── model_utils.py ← ML prediction & suggestion logic
├── district_aggregated.csv ← Real India Post dataset (754 districts)
├── requirements.txt ← Python dependencies
├── .gitignore ← Protects .env from being uploaded
├── README.md ← This file
└── .streamlit/
├── config.toml ← Streamlit theme & server settings
└── secrets.toml.example ← Template (copy → secrets.toml locally)
| Officer ID | Password | Name | Circle |
|---|---|---|---|
| IP2024KA001 | post@Karnataka | Ravi Kumar | Karnataka |
| IP2024MH001 | post@Maharashtra | Priya Desai | Maharashtra |
| IP2024TN001 | post@TamilNadu | Arjun Raj | Tamil Nadu |
| IP2024GJ001 | post@Gujarat | Neha Shah | Gujarat |
| DEMO | demo123 | Demo Officer | All India |
git clone https://github.com/YOUR_USERNAME/PostBot.git
cd PostBotpip install -r requirements.txtCreate a file called .env in the project folder:
GROQ_API_KEY=gsk_your_key_here
Get a free Groq key at → console.groq.com
streamlit run app.pyOpens at → http://localhost:8501
- Push this repo to GitHub
- Go to share.streamlit.io → New app
- Select your repo → branch:
main→ Main file:app.py - Click Deploy
- Go to Settings → Secrets and add:
GROQ_API_KEY = "gsk_your_actual_key_here"
| Component | Details |
|---|---|
| Prediction Model | CatBoost Regressor (R² = 0.9243) |
| Target Variable | district_delivery_rate (0–1) |
| Key Predictor | BO Ratio (Pearson r = +0.87) |
| AI Chatbot | Groq Llama 3.3 70B Versatile (Free — 14,400 req/day) |
| Dataset | 165,627 India Post offices → aggregated to 754 districts |
| Tier | Delivery Rate | BO Ratio |
|---|---|---|
| 🟢 High Performance | > 95% | ~90.5% |
| 🔵 Good Performance | 85–95% | ~86.1% |
| 🟡 Moderate Performance | 70–85% | ~83.0% |
| 🔴 Low Performance | < 70% | ~15.9% |
Author: Murali Manohara
Mentor: Mr. Aishwarya Sarda
Program: PGP in Data Science & GenAI, Great Lakes Institute of Management
This project is built for academic purposes as part of the Great Lakes PGP Capstone.