Intelligence layer
Nine real-world signal classes — continuously ingested, entity-linked, and assembled into a machine-readable model of the economy.
Source modules
Direct ingestion of SEC filings, quarterly earnings, and annual reports — structured into entity-linked fundamental intelligence records.
Signal ↗ Management sentiment shifts in 10-K risk sections detected ahead of guidance revisions.
Real-time streams across 40+ languages — parsed, de-duplicated, and structured into typed event candidates with entity linkage.
Signal ↗ Regional labor strikes flagged as company-level risk 18 hours before press coverage.
Precise mapping of 500 k+ corporate headquarters, stores, warehouses, and factories — continuously reconciled against company registries.
Signal ↗ Retail footprint overlap modelled during merger announcements for site-level risk.
Hyper-local historical and forecast data — integrated into consumption, logistics, and energy models at sub-municipality granularity.
Signal ↗ Unseasonal temperature deviations quantified against retailer sales-per-location baselines.
Real-time transit patterns and foot-traffic proxies — translated into physical-activity signals for industrial and consumer sectors.
Signal ↗ Highway congestion correlated with port-departure delays across Nordic logistics corridors.
Direct feeds from SCB, Eurostat, and the Federal Reserve — normalised, time-aligned, and linked to relevant entities and geographies.
Signal ↗ CPI acceleration in localised housing markets detected 6 weeks before national revision.
Comprehensive tracking of upcoming listings, dividend dates, earnings windows, and index rebalancing events.
Signal ↗ Quiet-period expirations of high-growth IPOs modelled against short-interest dynamics.
Propagation chains monitored through bills of lading and industrial logistics data — surfacing upstream disruption before earnings impact.
Signal ↗ Semiconductor delivery delays predicted through port-level congestion signals 3 weeks out.
Orbital-platform imagery processed into quantitative outputs — monitoring physical inventory levels and industrial output changes.
Signal ↗ Refinery utilisation estimated through thermal and visual change detection.
Intelligence pipeline
Every source object passes through a deterministic pipeline before becoming actionable intelligence. No black-box fusion. Every step is traceable.
Source Object
Raw signal ingested and typed
Normalization
Schema & time-alignment
Entity Linking
Connected to company graph
Context Modeling
Causal context assembled
Orion Output
LLM-ready intelligence object
Source governance
Corteqx evaluates external and proprietary streams by provenance, refresh cadence, entity coverage, and auditability before they enter the intelligence pipeline.
Explore Methodology