Developing sustainable land use strategies for rapidly urbanizing cities is a critical geographical challenge. While Transit-Oriented Development (TOD) offers a proven model, translating global best practices into locally relevant, actionable planning tools remains a significant barrier, particularly in developing countries. This paper introduces a transferable, four-step applied geographical framework to address this gap, using Phuoc Long metro station in Ho Chi Minh City, Vietnam, as a case study. The methodology first links the local station to an analogous global TOD typology using machine learning. Subsequently, a Random Forest model identifies the most critical, interpretable spatial features that define the typology. Finally, these data-driven insights are systematically integrated with local planning policies to propose context-specific TOD indicators. Applied to Phuoc Long Station, this process yielded a set of robust land use recommendations, including the conversion of existing industrial land (from 39.4% to 0%) and significant increases in mixed-use (to 15%) and green space (to 15%). The framework also produced key spatial technical indicators, such as a target population density of 150-190 persons/ha and the necessity of enhanced bus station density. The proposed methodology provides a quantitative framework for deriving place-based TOD indicators, offering a valuable tool for sustainable urban resource management in other developing cities.
Huynh et al. (Sun,) studied this question.