The staged Geopolitical Risk panel is designed as a persistent global command surface rather than a single composite table. It combines country-level pressure with a bounded stream of news and social discovery signals while keeping each evidence family separately inspectable.
Eight Lenses, One Geometry Contract
The candidate panel exposes Overview, Civil, Prices, Banking, Sovereign, Markets, Controls, and Evidence lenses. Each lens declares its own metric and coverage. Civil, price, banking, and evidence views are intended to preserve the difference between an observed zero and a missing field. Sovereign and market values should enter observed scoring only when their source and freshness contracts pass.
The global join uses the bundled ISO_A2_EH boundary field. Duplicate or invalid country rows are withheld rather than guessed. Neutral polygons mean no data; they are not a zero-risk category.
Source-Native Points Only
A news or social record becomes a point only when it carries defensible source-supplied coordinates. Coordinates marked inferred, estimated, geocoded, centroid-based, or otherwise derived are rejected even if the numbers are geographically valid. If the same record has a strict country association, its country polygon can pulse instead.
This policy prevents a common map failure: placing a precise-looking marker at a national capital or centroid when the source only names a country. Geometry caveats remain visible in point popups.
The Newest 20
Up to 200 classified discovery rows are retained within a 72-hour window. The newest 20 receive stable ranks and visual pulses. Ranking uses timestamp first and source score only as a tie-breaker. News and social channels remain separate, and the animation respects reduced-motion preferences.
These rows are discovery aids. A pulse does not confirm an event, severity, attribution, or causal relationship with the country score.
Replay Uses the Captured Clock
The candidate implementation reconstructs evidence windows, age labels, model timestamps, and newest-20 ranks from the snapshot capture time. It does not use the viewer's current clock. Live refresh controls are disabled during replay, and both direct news loading and late in-flight responses are blocked from mutating captured evidence.
Scoring Corrections
The candidate country-risk backend is intended to emit unsupported countries as unscored with a null score and null confidence. An unknown NR rating should not contribute a default risk floor. Additive score breakdowns reconstruct the displayed score exactly, including a visible cap adjustment when a result reaches 100.
- Synthetic CDS and debt baselines remain labeled context and are excluded from scoring.
- CDS observations require a valid, non-future as-of date within the accepted freshness window.
- Market rows require a trustworthy row or snapshot timestamp;
stale: falseis not enough. - Tax and fiscal diagnostics are not added again after they are represented inside sovereign-debt risk points.
- Registered or reachable source inventory does not count as risk evidence.
- SEC distress is labeled as a latest-filings sample when the feed is incomplete.
- CISA 30-day additions are counted from the complete available catalog before display sampling.
Candidate Verification Status
Reading the Map
When the map becomes available, start with the selected lens, not the polygon color alone. Open a country dossier to inspect source state, vintage, and evidence. Treat low-confidence and sparse rows as incomplete support, and treat discovery signals as prompts for source review rather than conclusions.
Related releaseConflict and Military OSINT Hardening->