The shapes are basically hand-made:

> And now a few words on methodology. As with all maps and statistics, these should be taken with several grains of salt. The goal here was to create a simple infographic for broad comparative purposes. To create the footprints themselves, I used satellite imagery to physically trace the boundaries of the built-up area of each city’s greater urban area. These footprints do not correspond to administrative boundaries. They are based purely on the divide between urban and rural land use. Which, at times, can be a very subjective task. I included low density suburban housing tracts within my urban footprints (hence the size of the US cities). I included dense built up areas which weren’t connected to the main contiguous urban area but were within its periphery (examples: Moscow, Frankfurt). I excluded rural areas, farmland, villages, or large urban parks. Obviously, simplification was necessary.

I love the idea, but the underlying data for the Tokyo region is wrong. The author seems to have used the population number for Tokyo Metropolis specifically (13.1M) but the geography for a much larger regional area which is closer to 30M people.

https://en.wikipedia.org/wiki/Greater_Tokyo_Area

The fuzzy notion of metropolitan area makes this less interesting that it might be.

An affordance with some crude tiers reflecting administrative layers and de facto ones, under user control or at least rendered out, would be a 10x improvement IMO.

As an SF resident the disjoint archipelago standing in for the Bay Area is wince-inducing.

Interesting. The shape of Amsterdam is super weird though. I wonder how the outline was made.