Tracking human behavioral patterns in cities can be used to determine urban delineations and urban land use, which has the potential to improve urban planning.
Urban areas are human settlements, typically cities, characterized by high population densities and built infrastructure. Urban areas need to be carefully planned, to ensure they are safe and sanitary. They are not self-sustaining but are dependent on an influx of essential resources.
Delineating urban areas is of great importance for planning and governance. Historically, this has been determined by establishing administrative boundaries and by surveying land use types. However, as urban areas have high populations, these approaches do not address how people interact with urban areas.
A newer approach to delineating urban areas, termed urban functional delineation, is now a focus of intense research. This approach is based entirely on analyzing how the population interacts with urban areas, independent of administrative boundaries or land use.
First and corresponding author Zahra Witsqa Maghfira, a graduate student in Hiroshima University's Transdisciplinary Science and Engineering Program , Graduate School of Advanced Science and Engineering , and co-authors at the Institut Teknologi Bandung, Indonesia, applied a social sensing approach to study urban functional delineation in Pekanbaru, Indonesia. Their research was published in the journal Computational Urban Science on July 7, 2025.
"We define an urban functional area as the spatial extent of concentrated urban functions," Maghfira explained. "We used Points of Interest (POI) data as a form of social sensing to capture behavioral patterns through locations that people frequently visit throughout the day. We analyzed this data using Kernel Density Estimation (KDE) and enhanced it with spatial autocorrelation, developing an urban delineation that reflects how urban spaces are used. This approach is especially valuable in cities like Pekanbaru, where conventional delineations may obscure the true dynamics of urban life."
"Our method identified urban spaces based on human activity instead of relying solely on built structures or satellite imagery," Maghfira elaborates. "The strong thematic and positional alignment with Sentinel-2 satellite data shows that our method detects urban presence reliably, but the distinct boundary patterns reveal something deeper: urban functionality is shaped by where people go and what they do, not just by what has been constructed. We argue that planning should embrace this behavioral perspective to better support dynamic, responsive urban development.
Using this approach, the researchers were able to identify five hotspots of urban activity in Pekanbaru. Comparison with land cover-, street network- and Sentinel-2 maps of the city revealed that the most-urbanized areas of the city determined by traditional methods and urban functional delineation were largely the same; however, the boundaries of the urban area differed sharply.
"Such disparity reveals a fascinating insight: activity-based models like ours do not replicate physical form but instead trace the rhythm and pulse of urban life," Maghfira says. "It listens to where a city is most alive. This difference is more than technical: it challenges how we interpret urban spaces and invites planners to recognize that where people go often matters more than what is physically there."
Future research will turn to applications of this method. Specifically, Maghfira is interested in using this approach to advise policymakers who make decisions concerning urban zoning and development. "We are developing a model that mimics the spatial logic used in formal zoning frameworks, aiming not only to replicate current patterns but to assess where in the spatial planning process such a behavior-based model can be most effectively applied. Ultimately, our goal is to create a decision-support tool that bridges real-world human activity and formal planning instruments, making zoning more responsive, adaptable, and reflective of actual urban dynamics," she concludes.
