State of the world today, scored 0–100. Higher = more geopolitical risk. Computed at 05:00 EST every day from 9 weighted components plus a bounded chaos term. See the math →
Initialising Leaflet · CartoDB tiles
State of the world today, scored 0–100. Higher = more geopolitical risk. Computed at 05:00 EST every day from 9 weighted components plus a bounded chaos term. See the math →
Click any chip to open the country drill-down. Watchlist subscribers get a daily email when a watched country shifts by |Δ| ≥ 10. Add countries via the + Watch this country button on any country panel, or paste an ISO code into the Settings panel.
See which countries moved most today — the 5 biggest GGRI increases and the 5 biggest decreases over the last 24 hours, with the events driving each move. Watchlist subscribers get a daily email when any watched country shifts by |Δ| ≥ 10.
Upgrade to Pro · $99/mo →Compare up to 3 countries on a single ten-category radar chart (security, political, civil unrest, crime, economic, regulatory, infrastructure, cyber, health, advisory). See exactly where each country’s risk concentrates and how they overlap or diverge in real time.
Upgrade to Pro · $99/mo →See every country in the world scored on the GGRI today, sortable by score, region, tier, or 24-hour delta. One-click watchlist toggle on every row — get a daily alert any time a watched country shifts by |Δ| ≥ 10. Free tier shows aggregated tier counts only; Pro unlocks all 193 individual scores and the full evidence trail.
Upgrade to Pro · $99/mo →| # | ISO | Country | Region | Tier | GGRI ▼ | Δ 24h | Watch |
|---|
AtlasRisks runs a continuous pipeline across thousands of public sources, normalized into a shared schema, scored against a transparent algorithm, and verified against known-hallucination patterns before it lights up the map.
GDELT, NewsAPI, RSS, ACLED, OFAC/EU/UK sanctions, World Bank/IMF, Reddit/Mastodon/Bluesky, public Telegram channels, ADS-B Exchange, AISHub/MarineTraffic. All public.
Political instability, security incidents, economic stress, social unrest, regulatory actions. Weighted sum with source-tier multiplier. Recency-decayed over 14-day half-life.
Cross-source corroboration check, image provenance (reverse + EXIF), LLM judge pass against structured ground truth, source-diversity weighting. Low-confidence items tagged.
Country scores, event feeds, military/maritime aggregates, source mix, and verification badges pushed to the map, the API, and subscribed briefings.
Every source AtlasRisks ingests carries an explicit credibility tier. Scoring is weighted — a Tier 1 government filing moves the score more than a Tier 4 anonymous post. Nothing is silently dropped; credibility is visible on every event.
AI-generated fake events and recycled-image stories are the biggest failure mode in public OSINT. AtlasRisks' verification loop runs every event through a fixed set of checks before it's allowed to influence a country score.
Events require ≥ 2 independent Tier 1–3 sources before influencing a country's score. A story reported only by Tier 4 shows in the feed but is flagged "single-source · unverified" and has a weight floor of 0.20×.
Reverse image search against Google / Bing / TinEye; EXIF metadata inspection; first-seen-date indexing. Images surfaced from unrelated dates or locations are rejected and the parent story is demoted.
Ten outlets citing the same original source count as one source, not ten. AtlasRisks chases citations back to the primary and scores against the root — so coverage volume doesn't amplify a single-point failure.
An LLM judge pass cross-references claimed actors, dates, and locations against structured datasets (ACLED, UCDP, government records). Contradictions trigger a manual review queue rather than publishing.
Images are run through open-source synthesis detectors (SDXL/Midjourney fingerprints, face-swap artifacts). Flagged content is marked machine-generated, provenance unclear and held out of scoring.
An event's impact on the score decays exponentially with a 14-day half-life. Recycled content — old stories re-surfacing as new — is detected by URL history and content hashing, and gets downweighted on re-emergence.
Card on file at signup, not charged for 15 days. The same zoomable Atlas you are looking at now, plus the full Event Timeline per country, watchlist alerts, and the GGRI Daily brief delivered to your inbox at 05:00 EST every day.
Start Free 15-Day Trial →Card on file · Cancel any time · GGRI methodology open