Left to right: Dr. Genevieve Giuliano (USC); Dr. Marlon Boarnet (USC); Yuquan (Wendy) Zhou, PhD Candidate (USC); Dr. Geoff Beoing, Zhou's disseration advisor (USC)

Yuquan (Wendy) Zhou, an Urban Planning & Development PhD candidate at USC's Sol Price School of Public Policy has been awarded the 40th Annual Charles M. Tiebout Prize in Regional Science at this year's Western Regional Science Association conference. This prize is awarded in honor of the outstanding contributions to regional science made by Charles M. Tiebout.

Zhou's paper titled, "Overestimating Essential Services Accessibility? Static vs. Dynamic Spatiotemporal Measures Using Individual Mobility Data in Los Angeles" was deemed the best graduate student paper in regional science.

Paper Abstract:
Accessibility metrics are widely used in transportation planning to determine where to locate services. However, many commonly used measures are static: they assume people travel from home, services are always available, and transit runs consistently. These assumptions can be far from reality, especially for people with constrained schedules and mobility. This study examined whether, and to what extent, static models overestimate people’s dynamic accessibility to essential services, and which service categories and population groups are most affected by this overestimation. Focusing on Los Angeles County, Zhou matched anonymized GPS trajectories of 1,286 individuals over one week with isochrones constructed using walkable, transit, and drivable networks and POIs with opening hours. Zhou compared people’s static, home-based accessibility with their dynamic, daytime-weighted accessibility that accounts for where they are and whether services are open at that time. Results show that static models systematically overestimate accessibility, especially for services with limited operating hours such as food and drink, health, and groceries. Transit-based accessibility shows the highest relative overestimation, due to additional constraints from operating schedules. To explain individual-level differences, Zhou used K-means clustering to identify three mobility typologies: non-movers, active local travelers, and occasional long-distance travelers and found that the overestimation is most severe for individuals with low mobility. By revealing how static models can bias real-world accessibility estimates and misidentify underserved populations, this paper contributes a scalable framework to measure dynamic people-based accessibility and support planners with better tools for informed planning decisions.