An international study led by Yatao Zhang of ETH Zurich and the University of Wisconsin, published in Nature Communications, shows that traffic congestion depends on more than road capacity. Zhang based the work on his doctoral thesis, completed in the autumn of 2025 at the Singapore‑ETH Centre. The team compared 30 major cities and focused on road traffic and congestion using vehicle movement data from Here Technologies with a five‑minute time resolution.
For a detailed city description the researchers relied mainly on OpenStreetMap. They collected information on the road network, traffic junctions, green spaces and the function of urban areas such as housing, shopping, sport, administration and education, and then correlated these features with congestion data. For Los Angeles the dataset included congestion values for over 18,000 road sections.
The team developed a new analytical method to study how urban features and traffic influence one another over time and to reveal cause‑and‑effect relationships that were previously hard to show. The results confirm a strong link between expansion of the road network and higher traffic levels, and they show that city form and building functions matter for traffic volumes. For example, a sprawling city tends to produce more traffic, clusters of leisure activities can increase weekend traffic, and mixed‑use areas with housing and work close together tend to shorten commutes and reduce traffic.
The comparison also highlights differences between cities: Singapore has clearly separated residential zones facing a centre with many service jobs, so changes in housing affect commuter flows, while in Zurich apartments are spread across the city and the link with commuter traffic is less pronounced. Martin Raubal, Professor of Geoinformation Engineering at ETH Zurich and a supervisor of the study, says the method can predict how a specific change—such as construction of a large shopping centre—will affect traffic in the medium term. The researchers note that further detailed analyses are needed before the method is used for local planning.
Difficult words
- congestion — a large number of vehicles slowing traffic flow
- network — a system of connected roads or routesroad network
- correlate — show a relationship between two thingscorrelated
- sprawling — growing outward with low density and spread
- commute — regular travel between home and workcommutes
- predict — say what will happen in the future
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- How might mixed-use development change commuting patterns in your city? Give examples.
- What risks could arise if a city responds to congestion only by expanding the road network?
- How should planners use predictive methods like this study while waiting for more detailed local analyses?
Related articles
Billions face cooling poverty as heat rises
A global study finds billions of people live in "cooling poverty" as hot spells become more frequent and intense. Researchers call for low-cost, coordinated policies across housing, water, health and cities because air conditioning alone is not enough.
