Scientific Abstract
Disaster mitigation across vulnerable Indian sub-regions requires high-fidelity predictive modeling combining historical geophysical disaster catalogues with historical meteorological telemetry. RESQ implements a medallion architecture (Bronze/Silver/Gold) ingesting multi-hazard records (cyclones, earthquakes, landslides, MODIS satellite fire data) coupled with 11-year Open-Meteo climate records to power dual-output XGBoost classifiers predicting incident type and hazard severity.
Methodology & Computational Modeling
11-year longitudinal backtesting across Indian states with multi-class disaster type and severity prediction.
MATHEMATICAL BASIS & CONTINUUM EQUATIONS:
Medallion ETL transformations, spatio-temporal lag convolutions, dual-output gradient-boosted decision trees (XGBoost).Observational Telemetry Datasets
This project ingests and assimilates open data streams from international registries:
Open-Meteo Climate Telemetry & MODIS Satellite Disaster Datasets
Validated Empirical Findings
- Automated medallion data pipeline (Bronze-to-Gold) normalizing multi-source disaster records.
- Dual-output XGBoost classification predicting hazard type and severity with localized weather enrichment.
- Integrated rescue coordination directory linking relief NGOs across 6 high-risk Indian states.
Investigators & Collaborators
Ayaan Saifi · Georbit Research
Matrika Regmi · Georbit Research
Qayad Ali · Georbit Research
Mehra Fatma · Georbit Research
Alisha Khan · Georbit Research
Nashra Zehra · Georbit Research
Khalid Mohammad Zaki · Georbit Research
Yahya Farooq · Georbit Research
System Access & Open Code
All project codebases and system architectures are openly accessible via their primary repositories and live deployments.