🏛️ System Architecture
Complete data flow from ingestion to output across all GMIIE subsystems
Data Pipeline
1 — Ingestion
Raw data from central banks, BIS, SWIFT, regulatory bodies, financial institutions, and open-source intelligence. Normalised into structured events with confidence scores.
2 — Analysis Engines
NLP-driven analysis layer extracting narrative clusters, measuring divergence between official statements and reality, scoring propaganda probability.
3 — 5-Ring Sovereign Engine
Five concentric analytical rings, each measuring a different dimension of sovereign monetary infrastructure evolution. The NIG score synthesises rhetoric vs. reality.
4 — Prediction Engine
Forward-looking engine combining all ring outputs to generate probabilistic scenarios, what-if simulations, and multi-dimensional societal impact assessments.
5 — Oracle Prediction Factory
5-layer probabilistic macro-asset forecast pipeline covering BTC, ETH, DXY, GOLD, and UST10Y. Ring signals feed into features. Regime-aware ensemble produces direction, return, and volatility forecasts with confidence intervals.
6 — Stabilization Layer
Calibration layer ensuring analytical integrity. Detects contradictory ring signals, dampens overconfident scores, enforces alert cooldown, enables human-in-the-loop review, and validates against historical crises.
7 — Output & Distribution
Multi-channel output layer delivering intelligence through interactive dashboards, real-time alerts, automated daily briefs, oracle forecast publications, and a comprehensive REST API.
Data Stores
PostgreSQL 16
Primary relational store — 34 model classes, events, ring scores, oracle forecasts, stabilization state
Neo4j 5
Graph database — 7 node labels, 7 edge types, infrastructure topology, dependency mapping
Redis 7
Cache & pub/sub — alert streaming, session cache, rate limiting, Celery broker