ARK
Frontier AI for Global Health Security
Autonomous Decision Intelligence
Synthesize global health, climate, socioeconomic, and geospatial signals into executive-grade decisions
Intelligence at the intersection of health, climate, socioeconomics, and place
Health Systems
Access, strength, and resilience indices aligned to clinical capacity.
Climate Exposure
Temperature, precipitation, and hazard anomalies with event rates.
Demographic Pressure
Population structure, dependency ratios, and growth acceleration.
Geospatial Context
Country and regional baselines to compare peer groups and hotspots.
From raw variables to executive decisions
ADI orchestrates derived features, composite indices, and model outputs into a structured decision fabric. Every card is traceable to the variable catalog with confidence indicators.
Decision cards built for executive clarity
Explore intelligence cards with real-time insights and actionable recommendations
Climate Health Risk Index
Derived from standardized domain indices with quantile bands.
Disease Surveillance Score
Derived from standardized domain indices with quantile bands.
Health System Resilience
Derived from standardized domain indices with quantile bands.
Navigate by continent, region, or country
ADI coverage aligns to the variable catalog and supports every country with available data. Use global filters to move from continental summaries to country detail.
Transparent, auditable decision engines
Every score, tier, and alert on ADI is computed by documented engines with published equations, parameter tables, and calibration dates.
Priority scoring with 45/25/30 need-capacity-urgency weighting, anomaly boost, and tier assignment calibrated against WHO HEPR.
Z-score severity classification (CRITICAL/SEVERE/HIGH/MODERATE) with crisis probability CDF and robust MAD-based detection.
Min-max normalization, arithmetic/geometric/weighted aggregation with documented justification per KPI.
Fresh/aging/stale classification with temporal coherence scoring across data sources.
Standard and robust (MAD) z-scores with change-point detection for structural breaks in time series.
Every page shows collapsible methodology panels with equations, parameters, and data source citations.
Deep dive