Revealing hidden insights in demographic data

A map of the world showing population growth from migration from 2000-2020. The scale is a color gradient, with orange showing that population change was driven by natural change and purple showing that population change was driven by migration.

The University of Michigan contributed to an analysis of more than 11,000 cities worldwide, revealing trends that are hidden in national averages. For example, population changes were more likely to be driven by migration in mega cities compared where smaller cities, where natural changes—births and deaths—were more likely to be responsible. Image credit: A. Zimmer et al. Nat. Cities. 2026 DOI: 10.1038/s44284-026-00447-7 (Used under a CC-BY license)

By Diana Setterberg | Montana State University News Service |

June 1, 2026

Contact: Matt Davenport, Michigan News

A new analysis of data from more than 10,000 cities worldwide provides governments and agencies with information to help evaluate cities' strengths and vulnerabilities, and determine how best to allocate resources and support populations.

“Our results show that national averages obscure substantial differences between cities and that globally consistent city-level demographic data can provide important insight for urban planning, climate adaptation and development,” said Nina Brooks, University of Michigan School for Environment and Sustainability (SEAS) assistant professor, who is an author of the study.

The team’s analysis, supported in part by NASA, revealed that, from 2000 to 2020:

Globally, 45% of urban population growth was due to migration.
Migration drove growth primarily in megacities, whereas natural population dynamics (births and deaths) were the dominant factors in smaller cities.
Smaller cities remained younger than larger cities, especially in Africa.
Developing cities, especially in Asia, Africa and Latin America, are changing more quickly than developed cities.
Urban dependency ratios, which are the number of children and older adults relative to the working-age population in cities, generally declined over the study period because working-age adults moved from rural to urban areas.

The team blended multiple datasets to get comprehensive data on both the demographics and the spatial extent of cities. To blend the datasets, the research team used the Tempest supercomputer at Montana State’s Research Cyberinfrastructure core facility.