In a single year, storms and floods can hit three quarters — or more — of an island nation's entire population. Global statistics erase this, because we count in absolute totals and encode by land area.
When the world maps a global crisis, it falls back on national area maps. They encode values through a country's surface area — handing visual weight to large states regardless of how many people are actually hit. The result is a systematic distortion.
On a world map they are barely more than a pixel. This is exactly where the front line of the climate crisis runs.
Measured in absolute numbers, the large states top the table of people affected. Recalculate the same events as a share of population, and the ranking flips completely. Looking only at absolute figures erases the smallest nations statistically — scroll on and watch it happen.
On the right-hand axis, some countries exceed 100 %. This is not a charting error but a property of the source: when several storms or floods hit the same population one after another within a single year, the UNDRR/Sendai records count the same people again for each event. The total can therefore exceed the number of inhabitants. Honest visualisation discloses this — those values are flagged.
On the absolute axis, Tuvalu, the Marshall Islands and Nauru sink into the bottom third — as if barely affected. Relative to population the picture inverts entirely: in Tuvalu 77 % of people were hit, in the Marshall Islands nominally over 100 %, in Nauru 46 % — figures no large state ever reaches. Mapping the world in absolute totals hides precisely those nations whose existence is at stake. The question is no longer where the most people suffer — but why these states in particular.
"Affected" is not a soft word here. When a storm flattens the coconut palms or saltwater poisons the soil, crop yield collapses in those same years — and with it food, income and the plain habitability of an island. Yield is the hardest proxy the data offers: it measures whether the physical basis of life still holds. You will find that curve per country in a moment.
Two linked views of the same twelve countries. On the left, causation — CO₂ per capita — against exposure. The finding is descriptive, not indignant: the states with the lowest per-capita emissions sit at the very top for exposure. On the right, the rank shift from absolute to per capita. Hover over a country in one view — it lights up in both at once.
For scale: average CO₂ emissions in Europe are roughly 6–7 tonnes per capita. · Correlation, not causation — the relationship is shown, not claimed as a cause. Bright lines = the three atoll nations in focus; dimmed = context.
Compress a whole country into a single figure and the place where the island is going under disappears — the national mean irons the extreme flat. And depending on how coarsely you cut the map into units, a different picture emerges. The same data, bundled differently, tells a different story. Terms: Ecological Fallacy · Modifiable Areal Unit Problem
How we measure ultimately decides who receives help.
Click to finishThose with little land area become invisible on maps, however existential the threat. Area bias · MAUP
Click for the next cardWithout relative context it conceals the catastrophe.
Click for the next cardAnd whoever holds it decides who stays in the dark.