115 laps at Thunderhill
A year of my own telemetry, against a theoretical-best floor.
7.9 billion
A broadsheet on where people actually live. Five continents pulled down through thirty-three countries into fifty cities, read as flow instead of as a map.
The question was how many people, and under which country
Population on a world map is a solved and slightly useless picture. The eye goes to landmass, and landmass is the one variable that has nothing to do with the question. Russia looks like the story. It is not the story.
So the sheet drops the map to a thumbnail and gives the center to an alluvial diagram: three stacked axes, continent to country to city, with a thread for every city. Read down, a thread answers a chain of questions at once. Which continent, which country inside it, which city inside that. The crossings in the middle are not noise. They are the diagram admitting that rank at one level does not survive to the next.
Ten cities per continent instead of the global top fifty
The fifty cities are not the fifty largest on Earth. They are the ten largest on each of five continents, a deliberate quota, and the decision that shapes everything downstream. A straight global top fifty would have been an Asia list with footnotes. Asia alone carries 160.8 million across its ten, more than four times North America's ten.
The quota costs accuracy about the world and buys legibility about the continents. It is why Philadelphia, at 1.57 million, shares an axis with Shanghai at 24.26 million. Fifteen times its size, and still North America's tenth city. That ratio is the argument, and it only shows up because both are on the sheet.
160,840,058 · 42.3%
70,169,541 · 18.5%
56,956,851 · 15.0%
55,479,685 · 14.6%
36,399,988 · 9.6%
Ten cities per continent, fifty in all, across thirty-three countries. 379,846,123 people. Bars are scaled to Asia.
One hue, used only for population
The sheet is black, white, and a single red. Type, rules, threads, axis labels and the locator map are all greyscale. The red appears exactly twice, and both times it is a population: the nested circles in THE NUMBERS, and the row of circles closing every thread on the city axis.
That is the whole color system, and it is the same rule the rest of this site runs on. The interface is greyscale and color is reserved for data. On a sheet this dense it is doing real work. With one hue in play, a red dot cannot mean anything except a number of people, so the bottom axis reads as a scale instead of as decoration.
The circles are area-scaled, which is why Shanghai does not simply obliterate the row. Against Philadelphia it is 15.5× the population but only 3.9× the width. Had population driven the radius instead, Shanghai would have read as roughly 240×. That is the classic way a bubble chart lies.
I checked my own encoding instead of trusting it
The area scaling is the load-bearing decision on this sheet, so I measured it rather than assuming the software got it right. Sampled against the five labeled circles, the encoding holds to within half a pixel on a 734-pixel base, and all fifty city circles carry the same constant.
That check is the habit I care about here. A bubble chart that lies looks exactly like one that does not, which is why the only way to know is to go back and measure. It is the same instinct that later turned into Color Safety Lens: if a visual decision can be verified with arithmetic, verify it.