May 22, 2026
Bookmark feature is for subscribers only. Subscribe Now to save your favorites.

How Does Land Use Affect Public Vehicle Speeds in Developing Cities?

FacebookXEmailCopy Link
Researchers leverage machine learning and tricycle GPS data to demonstrate how localized land use patterns directly drive traffic congestion in developing urban centers.

Researchers leverage machine learning and tricycle GPS data to demonstrate how localized land use patterns directly drive traffic congestion in developing urban centers.

Urban transport research has long depended on data-rich cities in developed countries. Traffic in developing cities follows a very different rhythm, however, according to researchers. In developing cities, data is often sparse, infrastructure is unevenly digitized, and roads are shared by a mix of transport modes including jeepneys, motorcycles, tricycles, buses, informal shuttles, and private vehicles.

A January 2024 study published in Cities, a peer-reviewed journal by Elsevier, helps address this gap. The research found that land use patterns, road speed limits, and time of day influence public vehicle speeds in developing urban areas.

“Many transport models are built around highly developed urban systems,” the researchers said," but mobility in developing cities is far more complex and informal, so imported assumptions do not always work.”

Using GPS and Machine Learning to Study Traffic

To better understand transport dynamics in low-resource urban settings, a research team led by National Academy of Science and Technology Academician and Asian Institute of Management Aboitiz School of Innovation, Technology, and Entrepreneurship Head Christopher Monterola used machine learning to analyze how land use affects mobility.

“We wanted to understand how the structure of a city itself shapes movement,” the researchers said. “Not just the roads, but the surrounding land use, the activity centers, and the way public transport interacts with them.”

The team focused on Cauayan City in Isabela, Philippines, a growing city with smart city ambitions but limited transport data infrastructure. Researchers worked with the local government to install GPS sensors on tricycles, one of the city’s most common forms of public transport.

“Tricycles are an important part of mobility in many Philippine cities, yet they are often invisible in transport research,” the team noted. “Using them as probe vehicles allowed us to capture everyday traffic behavior at street level.”

Using building footprints, satellite imagery, road network data, and GPS traces, the researchers trained machine learning models to predict one-minute average vehicle speeds across different parts of the city and at different times of day.

The models were then compared with baseline approaches based on OpenStreetMap and historical road speed averages.

What the Study Found

The findings revealed that land use plays a major role in shaping traffic flow. Vehicles moved more slowly near dense residential and commercial areas, while speeds increased in areas with fewer nearby buildings.

“Congestion is not caused only by the number of vehicles on the road,” the researchers wrote. “The surrounding urban environment matters a great deal.”

This content is available to our subscribers. Join our community to access premium content and elevate your business knowledge.

Already a subscriber? Login
OR
ANNUAL
1,000
per year
SEMI-ANNUAL
500
per six months
QUARTERLY
250
per three months
MONTHLY
100
per month

Frequently Asked Questions

Traffic in developing cities follows a much more complex and informal rhythm because datasets are often sparse and infrastructure is unevenly digitized. Furthermore, roads in these urban systems are shared by a diverse and unstructured mix of transport modes, including jeepneys, motorcycles, tricycles, buses, informal shuttles, and private vehicles, meaning imported transport models from developed countries do not always apply.

The research team, led by Christopher Monterola, worked closely with the local government of Cauayan City to install physical GPS sensors directly onto tricycles. Since tricycles are one of the city's most common forms of public transport but are typically invisible in transport research, using them as probe vehicles allowed the team to successfully capture everyday traffic behavior at the street level.

Vehicles move significantly slower near dense residential and commercial areas due to specific urban bottlenecks. Large malls slow traffic by attracting heavy pedestrian and vehicle activity, while smaller commercial establishments create bottlenecks when vehicles spill into the street due to limited parking, and residential areas contribute to slower speeds because of narrow roads and roadside parking.

The study revealed that traffic congestion is shaped not just by a vehicle's immediate surroundings, but by the wider neighborhood context. Machine learning models that analyzed land use across multiple distances, ranging from fifty to one thousand meters, performed significantly better, proving that congestion behaves like a neighborhood effect rather than an isolated street-segment issue.

By utilizing an interpretable machine learning technique called SHAP, the researchers identified that nearby residential areas and official road speed limits serve as the strongest overall predictors of vehicle speed. These primary factors are closely followed by the total number of road lanes, the specific time of day, and proximity to commercial establishments.

Rocky Teodoro

Rocky Teodoro

Writer

Rocky Teodoro is a writer and editor with 2 decades of experience. He has previously served as a senior manager for News and Research in S&P Global. He has also served as a managing editor for The Business Manual and a news editor for oil and gas portal Rigzone. In his editorial career, he also has stints as a technical writer, features writer, manuscript editor, and magazine contributor.

Learn straight from the top CEOs and business leaders. Access exclusive articles and videos.

Subscribe Now