🌡️ ClimateStock India — Climate-Finance Predictor

A full-stack, production-grade web application that correlates Indian regional climate data (temperature, heatwave events) with energy sector stock performance, featuring real-time XGBoost ML predictions, AI chatbot, and live PostgreSQL dashboards.

📸 Features at a Glance

Feature	Description
📊 Live Dashboard	Real-time temperature vs. stock price area chart, updating every 8 seconds
🤖 ML Prediction Engine	XGBoost model predicts energy stock price from climate inputs
💬 AI Chatbot (ClimateIQ)	Gemini-powered context-aware chatbot grounded in live DB data
📈 Insights Page	Statistical analysis: heatwave impact, price volatility, prediction trends
🌡️ Climate Events Donut	Full-table pie chart showing Normal Days vs. Heatwave Alerts
⚡ Auto DB Seeding	On first run, auto-creates schema and seeds from CSV — zero setup

---

🗂️ Project Structure

Market/
├── backend/                        # Node.js Express API server
│   ├── server.js                   # Main API server (all routes)
│   ├── predict.py                  # Python XGBoost inference script
│   ├── chat.py                     # Python Gemini AI chatbot script
│   ├── init_db.sql                 # PostgreSQL schema (tables definition)
│   ├── init_db.js                  # Manual DB initializer script
│   ├── seed_csv.js                 # Manual CSV seeder script
│   ├── list_models.js              # Utility to list available Gemini models
│   ├── package.json                # Node.js dependencies
│   └── .env                        # Environment variables (secrets)
│
├── frontend/                       # React + Vite frontend
│   ├── src/
│   │   ├── pages/
│   │   │   ├── Dashboard.jsx       # Market Overview page (charts + stats)
│   │   │   ├── Prediction.jsx      # ML Prediction input & result page
│   │   │   └── Insights.jsx        # Statistical insights dashboard
│   │   ├── components/
│   │   │   ├── ChatBot.jsx         # Floating AI chat widget
│   │   │   └── ChatBot.css         # Chatbot-specific styles
│   │   ├── App.jsx                 # Root app with routing & sidebar nav
│   │   ├── App.css                 # App-level layout styles
│   │   └── index.css               # Global CSS design system & variables
│   ├── index.html
│   └── package.json
│
├── final_processed_data.csv        # Historical climate + stock dataset
├── xgb_stock_prediction_model_raw.pkl  # Pre-trained XGBoost regressor (price prediction)
├── rf_classifier_20k_model.pkl     # Pre-trained Random Forest classifier (trend classification)
├── inspect_model.py                # Utility to inspect model features
└── README.md                       # This file

🛠️ Tech Stack

Backend

Technology	Purpose
Node.js + Express	REST API server
PostgreSQL	Persistent database for all historical and prediction data
Python 3	XGBoost inference (predict.py) and Gemini AI bridge (chat.py)
XGBoost	Pre-trained ML model for stock price prediction
Google Gemini API	Powers the ClimateIQ AI chat assistant
dotenv	Manages environment variables

Frontend

Technology	Purpose
React 18 + Vite	Modern SPA framework
Recharts	Area charts, pie/donut charts for data visualization
Axios	HTTP client for API calls
Lucide React	Icon library
Vanilla CSS	Custom CSS design system (no frameworks)

---

🗃️ Database Schema

The PostgreSQL database Stockmarket has three core tables:

companies

Column	Type	Description
id	SERIAL PK	Unique ID
name	TEXT	Company name (e.g., Reliance)
symbol	TEXT UNIQUE	Stock ticker symbol
sector	TEXT	Always Energy for this dataset

temperature_data

Column	Type	Description
id	SERIAL PK	Unique ID
date	DATE	Record date
region	TEXT	Region (default Unknown)
avg_temperature	NUMERIC	Temperature in °C
heatwave_flag	BOOLEAN	true if heatwave day

stock_data

Column	Type	Description
id	SERIAL PK	Unique ID
company_id	INT FK	References companies.id
date	DATE	Record date
open_price	NUMERIC	Open price (₹)
close_price	NUMERIC	Close price (₹)
high_price	NUMERIC	Day high (₹)
low_price	NUMERIC	Day low (₹)
volume	BIGINT	Trade volume

predictions

Column	Type	Description
id	SERIAL PK	Unique ID
company_id	INT FK	References companies.id
date	DATE	Prediction date
input_temperature	NUMERIC	Temperature used as input (°C)
predicted_price	NUMERIC	XGBoost predicted price (₹)
predicted_trend	TEXT	Bullish, Bearish, or Neutral
confidence_score	NUMERIC	Model confidence (0–100%)

---

🤖 ML Models — Dual-Model Architecture.

The prediction engine uses two complementary ML models working together:

1. XGBoost Regressor — Price Prediction

Model file: xgb_stock_prediction_model_raw.pkl

Trained on Indian energy sector historical data combining climate and financial features. Predicts the next-day stock closing price (₹).

2. Random Forest Classifier — Trend Classification

Model file: rf_classifier_20k_model.pkl

A 20,000-tree Random Forest classifier that predicts the market trend direction (Bullish / Bearish / Neutral) and provides probability-based confidence scores via predict_proba.

Input Features (15 total)

Feature	Description
Temperature	Regional average temperature (°C)
CDD	Cooling Degree Days (proxy: max(0, Temp - 25))
CDD_7	7-day CDD accumulator
Heatwave	Binary flag: 1 = heatwave day
Return_Lag1	1-day return lag
Return_Lag3	3-day return lag
Return_Lag5	5-day return lag
Temp_Anomaly	Deviation from 30°C baseline
CDD_Lag1	CDD 1-day lag
CDD_Lag3	CDD 3-day lag
MA_5	5-day moving average of close price
Volatility	Historical volatility
Demand_Trend	Energy demand trend index
Trend	Market trend signal
Volatility_Change	Change in volatility

Output

• Predicted Close Price (₹) — from XGBoost regressor
• Trend (Bullish/Bearish/Neutral) — from RF classifier
• Confidence Score (%) — from RF classifier's predict_proba

---

📡 API Endpoints

All endpoints are served on http://localhost:5000.

Method	Endpoint	Description
GET	/api/stocks	Fetch all companies
GET	/api/temperature	Latest 30 temperature records
GET	/api/dashboard	Time-offset merged climate + stock data (30 rows)
POST	/api/predict	Run XGBoost prediction with climate inputs
GET	/api/insights	Aggregated statistical insights from DB
GET	/api/climate-summary	Full-table heatwave/normal day counts
POST	/api/chat	Send message to ClimateIQ AI chatbot

POST /api/predict — Request Body

{
  "temperature": 38.5,
  "heatwave_flag": true,
  "company_id": 1
}

POST /api/predict — Response

{
  "prediction": 872.34,
  "trend": "Bullish",
  "confidence": 78.5,
  "returnPct": 2.45,
  "volatility": 0.0415
}

POST /api/chat — Request Body

{
  "message": "How do heatwaves affect energy stocks?",
  "history": []
}

---

⚙️ Environment Variables

Create a .env file inside the backend/ folder:

# PostgreSQL Database
DB_NAME=Stockmarket
DB_USER=postgres
DB_PASSWORD=your_password_here
DB_HOST=localhost
DB_PORT=5432

# Server
PORT=5000

# Google Gemini API Key (for AI chatbot)
GEMINI_API_KEY=your_gemini_api_key_here

# JWT (for future auth)
JWT_SECRET=your_secret_key_here

⚠️ Never commit your .env file to Git. Add it to .gitignore.

---

🚀 Getting Started

Prerequisites

Make sure the following are installed on your system:

• Node.js v18+
• Python 3.8+
• PostgreSQL v14+
• Python packages: xgboost, pandas, numpy, scikit-learn

1. Install Python Dependencies

pip install xgboost pandas numpy scikit-learn

2. Set Up the Database

Create the database in PostgreSQL:

CREATE DATABASE Stockmarket;

Then update your backend/.env file with the correct credentials.

3. Start the Backend

cd backend
npm install
node server.js

On first start, the server will automatically:

1. Connect to PostgreSQL
2. Create all tables (from init_db.sql)
3. Seed data from final_processed_data .csv

You should see:

✅ Connected to PostgreSQL Database successfully!
✅ Schema created successfully.
⌛ Inserting records (this may take a minute)...
✅ Successfully seeded XXXXX records!
Server running on port 5000

4. Start the Frontend

Open a new terminal:

cd frontend
npm install
npm run dev

Visit: http://localhost:5173

---

📊 Dataset — final_processed_data .csv

The dataset contains historical Indian energy sector climate-finance data.

Key Columns

Column	Description
Date	Trading date
Stock	Company ticker symbol
CLOSE	Closing stock price (₹)
Temperature	Regional temperature (°C)
Heatwave	Binary: 1 if heatwave day, 0 otherwise

The data covers multiple Indian energy companies including Reliance Industries, Tata Power, NTPC, Adani Green, and Power Grid Corp.

---

🧪 Inspecting the ML Model

Run the inspection utility to see model feature names and type:

python inspect_model.py

---

🌐 Application Pages

1. Dashboard (/)

• 3 Stat Cards: Avg Temperature, Latest Stock Price, Climate Risk Score
• Area Chart: Temperature (°C, left axis) vs. Stock Price (₹, right axis) — live 8s refresh
• Donut Chart: Climate Events split — Normal Days vs. Heatwave Alerts

2. Prediction (/predict)

• Input form: Temperature (°C), Heatwave toggle, Company selector
• Circular progress indicator for Confidence Score (color-coded: Green/Orange/Red)
• Displays: Predicted Price, Trend (Bullish/Bearish/Neutral), Return %, Volatility

3. Insights (/insights)

• Temperature statistics: Avg, Max, Min, Standard Deviation
• Stock price analysis: Heatwave-day vs. Normal-day averages, Volatility %
• Prediction history: Total predictions, Bullish/Bearish/Neutral distribution

4. AI Chatbot (ClimateIQ)

• Floating chat widget available on all pages
• Context-aware: grounded in live PostgreSQL data (avg temp, stock price, predictions)
• Powered by Google Gemini via Python bridge

---

🔧 Troubleshooting

Backend files showing red (D) in VS Code

Files are staged for deletion in Git. Run in the Market folder:

git restore --staged .

python not recognized

Make sure Python is added to your system PATH, or use python3:

python3 --version

Database connection error

• Verify PostgreSQL service is running
• Check DB_PASSWORD and DB_USER in backend/.env
• Ensure database Stockmarket exists:

  CREATE DATABASE Stockmarket;

Port 5000 already in use

Change the port in backend/.env:

PORT=5001

And update API URLs in frontend pages from 5000 to 5001.

Model file not found

Ensure both xgb_stock_prediction_model_raw.pkl and rf_classifier_20k_model.pkl are in the root Market/ directory (not inside backend/).

---

📁 .gitignore Recommendations

# Dependencies
node_modules/
__pycache__/
*.pyc

# Environment
.env

# Build output
dist/
build/

# Data & Models (optional, large files)
*.pkl
final_processed_data .csv

---

🤝 Contributing

1. Fork the repository
2. Create your feature branch: git checkout -b feature/your-feature
3. Commit your changes: git commit -m 'Add some feature'
4. Push to the branch: git push origin feature/your-feature
5. Open a Pull Request

---

📄 License

This project is for academic and research purposes.

---

👨‍💻 Author

Bhargav — Full-Stack Developer
ClimateStock India Predictor · Built with ❤️ using React, Node.js, Python & XGBoost