🛠 Status: Prototype (v1.0) 💧 Sector: Water Supply 💻 Tech Stack: Python | AI | ML | Geodata 📍 Region: Haryana, India (East-West Analysis)
This project is a technical prototype developed for the Haryana AI Sandbox (HAIDP) initiative. It transforms raw NASA satellite data into an actionable "Risk Rating" system for Haryana’s groundwater resources.
By analyzing mass anomalies, we identify the "Water Burn Rate"—the velocity at which the state is depleting its underground water capital.
Our analysis reveals a significant disparity between regional depletion rates:
- East Zone (Industrial/Urban): Highest depletion velocity due to NCR urbanization and intensive cropping.
- West Zone (Agricultural): Moderate depletion, buffered by surface canal systems.
- State-Wide Burn Rate: -0.4638 cm/month (~5.5 cm/year).
- Data Source: NASA GRACE-FO (Gravity Recovery and Climate Experiment) Level-3 Mascon CRI-Filtered RL06.3.
- Processing Engine: Python (
xarray,scipy,numpy) for spatial clipping and linear regression analysis. - Visualization: Interactive Folium map integrated into a Streamlit dashboard.
- Scientific Standard: Built using methodologies learned through IIRS/ISRO Geodata Analysis certification.
- ML Downscaling: Utilizing Random Forest models to increase resolution from 50km to 5km using Sentinel-2 (NDVI) as a proxy.
- Ground-Truthing: Ingesting in-situ data from Haryana's piezometric observation wells.
- Predictive Analytics: Implementing LSTM (Long Short-Term Memory) networks to forecast "Day Zero" for all 22 districts.
Srinivasta
- 20+ Years in Market Risk & Analytics
- 4 Years Hands-on AI/ML Development
- Certified in Geodata Analysis (IIRS/ISRO)
- LinkedIn Profile | Portfolio App
Developed for the Haryana AI Sandbox Launch - June 1, 2026