About.

What this is

A globe choropleth of country-level causes of death, drawn from the IHME Global Burden of Disease 2023 dataset. Click any country and the side panel lists what tends to kill its residents - either by annual death rate per 100,000 people, or by the lifetime probability that a person will eventually die of that cause.

The idea took shape one evening while I was watching 72 Dangerous Animals - the Netflix series that walks through a region at a time (Asia, Australia, Latin America) and ranks its wildlife by how thoroughly it will ruin your day. I started wondering what the equivalent map for the actual leading causes of death would look like at country resolution, and this atlas is what came out of a few evenings spent answering that question.

There are no ads, no donations, and no paid tier.

How to read it

1. Click a country to open its profile. Causes are ranked by the metric you’ve picked - either deaths per 100,000 in 2023, or lifetime probability of death from that cause.
2. Toggle the metric in the header. Both the choropleth and the country profile re-rank accordingly; the underlying GBD record is the same.
3. Pick a risk category in the sidebar to recolour the globe. The choropleth then shades countries by the aggregate value within that category, using the active metric.
4. Drag the globe to rotate it. Colours respect your system colour scheme.

The two metrics

Deaths per 100,000is the GBD’s age-standardised annual death rate: out of every 100,000 people in the country in 2023, how many died of that cause that year. It is the conventional public-health yardstick and is comparable across countries with different population sizes.

Probability of deathis IHME’s lifetime cause-specific probability: the share of a cohort that will eventually die of that cause, assuming current cause-specific mortality patterns hold. A 15 % figure for ischaemic heart disease means roughly one in seven people in that country, over a full lifetime, will die of it.

The two views agree on the broad shape of mortality but disagree on the margins. Lifetime probability flattens fast-killing acute causes a little and compounds slow chronic ones; the annual rate does the opposite. Switching between them is the point of the toggle.

That disagreement is also why the globe sometimes recolours when you switch metrics. The choropleth doesn’t shade by absolute value; it shades by each country’s rank within the selected category. Two countries can have a similar annual death rate from, say, heart disease while sitting far apart on lifetime probability - because PoD also reflects how long people in that country actually live and what other causes compete for them. A country with a younger population can rank low on lifetime probability of a chronic cause even when its annual rate is middling; an older population pushes the opposite way. When a country’s rank moves, its colour stop on the ramp moves with it.

Risk categories

The GBD publishes hundreds of fine-grained causes at its most detailed level - “Ischaemic heart disease,” “Road injuries,” “Self-harm by firearm,” and so on. This atlas rolls them up into thirteen broader risk categories - heart, cancer, infection, neurological, respiratory, diabetes, maternal, roads, accidents, violence, self-harm, substance use, and natural events - so the sidebar and choropleth stay legible at a glance. Causes that don’t fit a category are still listed in the country profile but excluded from the choropleth totals.

Sources & attribution

Country mortality rates come from the Global Burden of Disease Results Tool maintained by the Institute for Health Metrics and Evaluation (IHME) at the University of Washington. I use the 2023 estimates, published 2024, for two slices - “Deaths · Both sexes · All ages · Rate · all most detailed causes” and the matching “Probability of death, by cause” cube at the same demographic detail. Any error in translation is mine, not the GBD’s.

The IHME data is provided under their Free-of-Charge Non-Commercial User Agreement, which is why this atlas carries no advertising and asks nothing of you.

What it can’t tell you

Cause-of-death estimates at the country level are a model, not a census. Smaller countries carry wider uncertainty intervals, and a few are missing from the GBD slice entirely; where that happens the country renders blank.

A note on borders

The country shapes and names you see on the globe come from Natural Earth’s public dataset, via world-atlas. I did not draw them, do not edit them, and take no position on them. They reflect Natural Earth’s own editorial choices about disputed territories, occupied regions, and contested frontiers - choices that will not match every reader’s view of where lines should fall, particularly where those lines were redrawn by force. If a country is missing, mis-shaped, or labelled in a way you disagree with, the geometry is upstream of this atlas; please take it up with Natural Earth, not with me.

Colophon

Country geometry: Natural Earth, via world-atlas. Projection and topology handling: d3-geo and TopoJSON.

Built by Wojciech Sikora - Website · GitHub · LinkedIn.