About
How this works.
One score per country, built the same way for all 220 of them, from data anyone can check. Here's what goes into it — and, just as importantly, what it can't tell you.
The short version
Every country gets a score from 0 to 100. Higher means more risk. It's built from seven things that actually go wrong for travellers, each weighted by how much it matters:
Official advisories
25%How strongly the US, UK, Dutch, Canadian and German governments currently warn against going. Averaged, so no single country's politics drives it.
Source: US State Dept, UK FCDO, Netherlands MFA, Global Affairs Canada, German Federal Foreign Office
Conflict & terrorism
15%Armed conflict and political violence — how much fighting, terrorism and unrest the country is exposed to.
Source: INFORM Risk 2026 (European Commission)
Crime
15%Violent crime levels and how reliably the law is actually enforced. Based on murder rates, because that is the one crime figure countries cannot quietly under-report.
Source: UNODC via World Bank; World Bank governance indicators
Health & medical
15%How likely you are to get ill, and how good the care is if you do. Covers disease risk, malaria, water quality and hospitals.
Source: INFORM Risk; World Bank health indicators
Natural hazards
10%Exposure to earthquakes, floods, cyclones, drought and tsunamis. High here doesn't mean don't go — Japan scores badly and is very safe — it means check the season.
Source: INFORM Risk 2026 (European Commission)
Political stability
10%How likely unrest, a coup or sudden instability is to disrupt a trip — not whether you like the government.
Source: World Bank Worldwide Governance Indicators
Road & transport
10%Road deaths per 100,000 people. The risk travellers most underestimate, and the one that actually kills tourists.
Source: WHO via World Bank
Why government advisories count most
They're the only input that updates within hours of something happening, and the only one with a direct effect on your money: travel insurance is usually void in places your own government tells you not to go. We average five governments — the US, UK, Netherlands, Canada and Germany — so that no single country's foreign policy steers the score.
One rule we apply by hand
Averaging has an odd side effect: a country can have a war on and still land in the middle of the scale, because its roads and hospitals are fine. That's technically correct and completely useless. So when governments are near-unanimous that you shouldn't go, we push the score up to match. It affects 12 countries right now, and those pages say so.
What it can't tell you
Worth knowing before you lean on any of this.
- It doesn't know who you are. Risk isn't the same for a solo woman, a family, an LGBTQ+ traveller, or someone about to rent a scooter. That's why country pages show women's safety and LGBTQ+ legal status separately instead of burying them in one number.
- It can't see pickpockets. No country reliably counts petty theft, so the thing most likely to happen to you on a city break is the thing with the worst data. That's why the scam guides are written by hand.
- A country average hides everything. Mexico is one number. Mexico City and Sinaloa are not one place. Always read the map on the country page, not just the headline.
- Reported crime reflects who reports it. Countries where people trust the police record more crime. We lean on murder rates because that's the figure hardest to hide.
- Some of it is a year or two old. Advisories refresh every time we rebuild; health and economic figures are published annually. Each number on a country page shows its year.
Where the numbers come from
Everything is public and free. Nothing here is behind a paywall or a licence that stops you checking it yourself.
- World Bank Open Data — governance, homicide, road deaths, health and economic indicators
- INFORM Risk Index 2026 (European Commission JRC) — natural hazard, epidemic and conflict exposure, coping capacity
- UK Foreign, Commonwealth & Development Office — travel advice and regional warnings
- Netherlands Ministry of Foreign Affairs — colour-coded travel advice, including region-level zones
- US Department of State — travel advisory levels 1-4
- Natural Earth — country and admin-1 boundaries
- mledoze/countries — currencies, languages, dialling codes, capitals
- passport-index-dataset — visa requirements by passport
- Frankfurter / European Central Bank — exchange rates
A few countries — mostly small islands and territories — aren't covered by the global datasets. Rather than leave gaps that would flatter them, we estimate those from their neighbours and label it on the page. 15 countries are marked provisional for this reason.
Independence
Some links on this site are affiliate links and we may earn a commission from them. They're added after the scores are calculated and have no effect on any number. No country, tourism board or company can pay to change a score or remove a warning.
Corrections
If something here is wrong — a scam that doesn't happen any more, an area that's changed, a figure that misreads its source — please write to travelriskatlas@proton.me. Local knowledge beats a dataset, and the hand-written parts are the most likely to go stale.
Corrections that come with a source get made fastest, but a plain "this is not true any more" is genuinely useful too.
Using the data
The index is free to reuse with credit and a link. If you're a journalist or researcher and want the full dataset behind it — every score, every pillar, and a flag showing which figures are estimated — get in touch and we'll share it.
Current version: ARS v1.0, August 2026. First public release.