🌍 AetherIQ — Surface AQI Prediction & HCHO Hotspot Detection

ISRO Bharatiya Antariksh Hackathon 2026

AetherIQ is an end-to-end geospatial data acquisition pipeline that automates the collection of multi-source Earth Observation datasets for Surface Air Quality Index (AQI) Prediction and Formaldehyde (HCHO) Hotspot Detection across India using Google Earth Engine (GEE).

---

🚀 Features

• Automated Sentinel-5P trace gas downloads (HCHO, NO₂, SO₂, CO, O₃)
• ERA5-Land meteorological data acquisition
• MODIS & VIIRS active fire detection
• WorldPop population density
• SRTM elevation
• Automatic metadata generation
• SHA-256 checksum verification
• Resume interrupted downloads
• Retry mechanism with exponential backoff
• Download verification
• Production-ready modular architecture

---

🛠 Tech Stack

• Python 3.10+
• Google Earth Engine (GEE)
• geedim
• Rasterio
• NumPy
• Pandas
• tqdm

---

📦 Installation

# Clone the repository
git clone https://github.com/shingneshubham/AetherIQ.git

cd AetherIQ

# Create virtual environment
python -m venv .venv

# Activate (Windows)
.venv\Scripts\activate

# Activate (Linux/macOS)
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Authenticate Google Earth Engine
earthengine authenticate

---

📁 Project Structure

AetherIQ/
│
├── data/
│   └── raw/
│       ├── sentinel5p/
│       ├── era5/
│       ├── modis_fire/
│       ├── viirs_fire/
│       ├── worldpop/
│       ├── srtm/
│       ├── metadata.csv
│       └── checksums.csv
│
├── logs/
│
├── scripts/
│   ├── config.py
│   ├── utils.py
│   ├── download_sentinel5p.py
│   ├── download_era5.py
│   ├── download_modis_fire.py
│   ├── download_viirs_fire.py
│   ├── download_static.py
│   └── run_all.py
│
├── requirements.txt
├── README.md
└── .gitignore

---

🛰 Data Sources

Dataset	Source
Sentinel-5P	Google Earth Engine
ERA5-Land	Google Earth Engine
MODIS Fire	Google Earth Engine
VIIRS Fire	Google Earth Engine
WorldPop	Google Earth Engine
SRTM	Google Earth Engine

---

🔄 Pipeline Workflow

Sentinel-5P
      │
ERA5-Land
      │
MODIS Fire
      │
VIIRS Fire
      │
WorldPop
      │
SRTM
      │
────────────────────
Data Validation
      │
Metadata Logging
      │
SHA-256 Checksums
      │
GeoTIFF Dataset
      │
Preprocessing
      │
Surface AQI Prediction
      │
HCHO Hotspot Detection

---

▶ Running the Pipeline

Run Everything

python scripts/run_all.py

Run Individual Dataset

python scripts/run_all.py --dataset sentinel5p

python scripts/run_all.py --dataset era5

python scripts/run_all.py --dataset modis_fire

python scripts/run_all.py --dataset viirs_fire

python scripts/run_all.py --dataset static

---

📊 Dataset Statistics (2021–2024)

Dataset	Files
Sentinel-5P	240
ERA5-Land	48
MODIS Fire	48
VIIRS Fire	48
WorldPop	1
SRTM	1

Total Generated: 386 GeoTIFF files

---

✅ Production Features

• Direct-to-disk downloads using geedim
• Resume interrupted downloads
• Retry mechanism with exponential backoff
• Download verification
• SHA-256 checksum generation
• Metadata logging
• Per-dataset logging
• Fault-tolerant execution
• Modular architecture
• Command-line interface

---

📌 Manual Datasets

The following datasets are not available through Google Earth Engine and must be downloaded manually:

• CPCB AQI Ground Observations
• INSAT-3D Aerosol Optical Depth (AOD)

---

🔮 Future Work

• CPCB AQI preprocessing
• INSAT-3D integration
• Feature engineering
• Machine Learning model development
• Surface AQI prediction
• HCHO hotspot detection
• Interactive visualization dashboard
• Web deployment

---

🙏 Acknowledgements

• Google Earth Engine
• European Space Agency (Sentinel-5P)
• ECMWF
• NASA
• WorldPop
• ISRO Bharatiya Antariksh Hackathon 2026

---

📄 License

This project has been developed for the ISRO Bharatiya Antariksh Hackathon 2026.

---

👨‍💻 Author

Shubham Shingne

B.Tech CSE (Data Science)
Lovely Professional University

GitHub: https://github.com/shingneshubham