An interactive data analytics platform to analyze, predict, and visualize natural disaster relief budget allocations across Indian states — powered by ML algorithms.
🔗 Live Demo: natural-disaster-analyser-v0.vercel.app
This platform provides a comprehensive view of India's State Disaster Response Fund (SDRF) allocations from 2018–2025. It combines real disaster data with machine learning models to help analyze budget trends, predict future allocations, classify state-level risk, and cluster states by disaster activity.
Key stats covered:
- ₹1,52,700 Cr+ total SDRF budget across all states
- 67,359 deaths recorded across disasters (2020–25)
- 159 disaster events across 33 states/UTs
- Disaster types: Floods, Storms, Earthquakes, Extreme Temperature, Glacial Outbursts, Mass Movements
- KPI cards: Total budget, most affected state, total deaths, total events
- Bar chart: Top 10 states by SDRF allocation
- Pie chart: Disaster type distribution with percentages
- Line chart: Budget allocation & release trends over years
- State overview table: Events, deaths, people affected, seismic zone, and risk level
- Algorithm: Simple Linear Regression trained on 7 years of SDRF data (2018–2025)
- Select any state + target year → get predicted budget allocation
- Historical vs. predicted trend chart per state
- State comparison table: 2025 vs 2026 predicted budgets with % change
- Algorithm: Random Forest Classifier (5 decision trees, majority voting)
- Classifies all 33 states/UTs as High / Medium / Low risk
- Feature importance breakdown: Severity Score (30%), Disaster Frequency (25%), Budget Usage (15%), Seismic Risk (12%), Flood Risk (10%), Drought Risk (8%)
- Interactive risk heatmap of India
- Confidence scores and primary risk factors per state
- Algorithm: K-Means Clustering (3 clusters)
- Groups states by disaster frequency + SDRF budget allocation
- Clusters: Low Activity (18 states) · Moderate Activity (14 states) · High Activity (1 state)
- Scatter plot visualization with outlier detection
- Cluster insights for resource optimization and policy planning
- Upload custom disaster datasets
- Supports CSV format for extended analysis
| Layer | Technology |
|---|---|
| Frontend | Next.js, React.js, Tailwind CSS |
| Charts & Visualization | Recharts / Chart.js |
| ML Models | Linear Regression, Random Forest, K-Means (client-side) |
| Deployment | Vercel |
| Route | Description |
|---|---|
/ |
Main dashboard with KPIs, charts, and state overview |
/prediction |
Budget prediction using linear regression |
/risk |
ML-powered risk classification with heatmap |
/clustering |
K-Means clustering of states |
/upload |
Custom dataset upload |
- SDRF allocations: 2018–2025 (state-wise)
- Disaster events: 2020–2025 (159 events across 33 states/UTs)
- Attributes: Event count, deaths, people affected, seismic zone, flood/drought risk, budget allocation & release
# Clone the repository
git clone https://github.com/your-username/natural-disaster-analyser.git
# Navigate to project directory
cd natural-disaster-analyser
# Install dependencies
npm install
# Run the development server
npm run devOpen http://localhost:3000 in your browser.
- Live NDMA/government data API integration
- District-level granularity
- Time-series forecasting (ARIMA / LSTM)
- Multi-year custom range predictions
- PDF report export
- Mobile-optimized heatmap
This project is built for educational, portfolio, and research purposes.
Built to make disaster data accessible and actionable 🇮🇳