Overview
- A multi‑institutional team led by Zhe Zhu published a paper in PNAS on June 8 that introduces the “Urban Pulse,” a metric built from decades of NASA Harmonized Landsat and Sentinel‑2 (HLS) imagery.
- The study pairs HLS time‑series with a deep‑learning and time‑series method called CAPES to detect neighborhood‑level construction, repairs, demolition, and expansion into green spaces.
- Researchers applied the method to six cities—Seattle, Shenzhen, Lagos, Mumbai, Dubai, and Mexico City—and found three consistent patterns: development is spiky, follows boom‑and‑rest cycles, and is asynchronous across neighborhoods.
- The data captured a synchronized global slowdown in construction at the start of the COVID‑19 pandemic and revealed unequal recoveries across cities, with Shenzhen rebounding rapidly while Mumbai and Mexico City showed muted responses.
- Authors say the Urban Pulse could become an early‑warning diagnostic for planners, policymakers, and the public, but the work so far is a research demonstration and does not yet represent a standardized public service or fully deployed policy tool.