Urban Pulse: an EKG for cities from space

An EKG doesn’t show you what a heart looks like. Instead, it reveals what it is doing in real time, beat by beat.

Researchers have now created a comparable readout for cities, which they have named “urban pulse”.

Using high-frequency satellite imagery, they have produced a framework that follows the metabolic tempo of urban development as it unfolds - recording not only where cities expand, but also when it happens, how quickly it moves, and the pattern it takes.

The research was led by Zhe Zhu, Director of the Global Environmental Remote Sensing Laboratory at the University of Connecticut.

What was missing in urban data

For years, urban data has suffered from a time-lag. Planners and researchers have largely depended on infrequent surveys, rolled-up statistics, and snapshots separated by years.

That approach works for spotting big-picture shifts, but it is far less effective for detecting issues early or for understanding the cadence of how neighbourhoods truly evolve over time.

“For decades, we had just been capturing the outcome of urbanization – a house that’s been built, or a road expansion,” Zhu said. “But you don’t really see the dynamics within an urban area.”

The Urban Pulse framework tackles this gap by drawing on decades of dense time-series observations from NASA’s Harmonized Landsat and Sentinel-2 satellite datasets.

A deep-learning and time-series analysis technique known as CAPES was created by former postdoctoral researcher Ji Won Suh, now a tenure-track Assistant Professor at the University of Victoria in Canada.

CAPES analyses this imagery to pick up physical change at neighbourhood scale: fresh construction, demolition, infrastructure works, and encroachment into green space.

The outcome is far nearer to a continuous log than a handful of isolated snapshots.

Three underlying rhythms

To test the framework, the team applied it across six deliberately contrasting cities: Seattle, Shenzhen, Lagos, Mumbai, Dubai, and Mexico City. Despite their differences, all six displayed the same three fundamental patterns.

City growth is not a steady glide. Instead, it arrives in intense, episodic bursts - abrupt surges in building activity followed by calmer periods.

It is also cyclical, with neighbourhoods moving through phases of expansion and dormancy that do not map neatly on to any predictable seasonal schedule.

Urban growth is asynchronous, too: separate neighbourhoods can sit at very different points in the same city pulse at the very same time.

That apparent lack of coordination is not, in fact, a defect. It helps prevent a city’s infrastructure and labour markets from being overloaded all at once.

A city that tried to develop everywhere simultaneously would strain under its own momentum.

A cardiac arrest, then unequal recovery

The clearest illustration of the framework’s value came with COVID-19. When the pandemic arrived, the satellite record captured something resembling a sudden, synchronised cardiac arrest.

Construction activity fell steeply across cities around the world, at roughly the same moment. The next phase, however, differed dramatically from place to place.

Shenzhen experienced a pronounced drop followed by a fast, policy-driven rebound. Mumbai and Mexico City came back in other ways: more slowly and with greater unevenness, with the rhythm edging towards normal only gradually and not fully.

“It’s like in human beings,” Zhu noted. “When you get a disease, it’s not going to show up exactly the same in different people.”

That diversity of response is precisely what makes the tool informative. If every city reacted in the same way to an identical shock, there would be far less to learn from comparisons.

An early warning system

In the near term, the most obvious use is as a diagnostic aid for policymakers. Urban decline usually becomes visible only once it is far advanced - shuttered shops, falling footfall, and crumbling infrastructure.

By the time those symptoms appear, the underlying process has often been running for years. Acting at that point is slower, more costly, and less likely to succeed.

A system that follows neighbourhood-level construction rhythms in near-real time could surface early indicators before they accumulate into something much harder to reverse. It would offer a quiet form of warning that today’s data collection methods simply do not provide.

It could reveal unexpected dormancy in an area that would normally be active, or a pulse that begins to slow before anything looks wrong to the human eye.

The intended audience, though, is not confined to governments. The researchers aim to make the data openly available and practical for anyone trying to understand a city at street level.

“This is going to be a very impactful tool influencing not only top-down policy decisions from governments but also bottom-up decisions from everyday people navigating their cities,” Zhu said.

Cities have always had rhythms. They were just not legible before - at least not from space.