Density, rent, transport time — the three variables that define urban life, measured consistently across 192 cities from Tokyo to Lagos, and revealing patterns that upend conventional wisdom about what makes cities livable.
Every city is a machine for concentrating people, but no two machines work the same way. Tokyo packs 14 million people into a metropolitan area where average commute times by public transit are shorter than in Houston, a city one-tenth its size. Paris offers rents that — adjusted for density — are lower per accessible job than those in sprawling Atlanta. Mumbai achieves population densities that would be physically impossible under American zoning codes. These comparisons are not anecdotes: they are measurable, gridded, and now available in a single dataset covering 192 cities worldwide.
The dataset's architecture is deceptively simple. Each city is divided into grid cells. For each cell, four variables are recorded: population density, average rent, average real estate price, and travel time to the city center by both public and private transport. This uniformity is what makes the dataset powerful. Because every city is measured on the same grid with the same variables, researchers can ask questions that were previously unanswerable: How does the rent gradient — the rate at which rents decline with distance from the center — differ between monocentric European cities and polycentric American ones? How much does public transit access capitalize into housing prices? Do denser cities have shorter or longer commutes?
The answers challenge simplistic narratives. Density does not automatically mean congestion — several of the densest cities in the dataset have the shortest commute times, because density makes public transit viable. Nor does sprawl automatically mean affordability — low-density cities often have lower housing costs per square meter but higher total housing-plus-transport costs because residents must own cars and drive long distances. The dataset provides the empirical foundation for a more nuanced understanding of urban form, one that moves beyond the density-is-good-or-bad binary and toward quantifying the actual tradeoffs that different urban structures impose on their residents.
Average one-way commute time (minutes) across city types, showing that denser cities do not necessarily have longer commutes
Median rent as percentage of median household income, comparing compact vs sprawling metropolitan areas
| metric | this dataset | alternative | verdict |
|---|---|---|---|
| City Coverage | 192 cities worldwide | 50-80 cities (most urban datasets) | 2-4x broader |
| Variables per Cell | Density, rent, price, commute time | Density only (most gridded products) | Multi-dimensional |
| Transport Modes | Public and private, separately | Combined or private only | Modal comparison |
| Spatial Format | Gridded (consistent resolution) | Administrative boundaries (census) | Cross-city comparable |
| License | CC BY 4.0 | Restricted (many urban databases) | Fully open |
| Real Estate Data | Rent and purchase price | Purchase price only | Captures rental market |
The dataset provides empirical evidence for density-transit feedback loops: cities that invest in transit can support higher density, which in turn makes transit more viable, creating a virtuous cycle that reduces per-capita transport costs.
Comparing 192 cities reveals that housing affordability is not solely a supply problem — transport costs and urban form determine whether nominally cheap peripheral housing is actually affordable when total living costs are considered.
Urban form determines per-capita emissions. The dataset enables quantification of how much emissions reduction could be achieved by shifting growth patterns from sprawling to compact forms across rapidly urbanizing cities in the Global South.
Share this story