Authors: Roberto Ponce-Lopez*, Instituto Tecnologico de Estudios Superiores de Monterrey
Topics: Spatial Analysis & Modeling, Geographic Information Science and Systems, Urban and Regional Planning
Keywords: machine learning, unsupervised learning, clustering, non-work destinations, built-environment, central place theory
Session Type: Paper
Start / End Time: 9:55 AM / 11:35 AM
Room: Washington 6, Marriott, Exhibition Level
Presentation File: No File Uploaded
This paper describes a data-driven method to identify commercial patches at the interior of popular non-work destinations, and to characterize them through the spatial structure of Google Place API open data in Singapore. Our goal is to build typologies of spaces by the spatial structure of retail. For instance, to be able to recognize and classify a strip of restaurants and a small shopping mall. The attractiveness of a place over non-work activities lies in the composition of differentiated spaces catering to a diversity of people and activities. We want to define the character of a neighborhood by the type of commercial patches that it encloses. The methodology uses a combination of unsupervised machine learning techniques. First, a DBSCAN algorithm clusters establishments from Google Place API to identify spatial patches of points. Second, an algorithm computes several measures to represent the geometrical topology of the patch, the diversity and density of the establishments contained, and the spatial dependency of establishments on the presence of anchor shopping malls. Large datasets, such as the Google Place, demonstrate their potential to capture measures of the built environment, which allows us to characterize the built environment and place attraction with a high level of spatial detail. Beyond the methodological contribution, a substantial finding is the discovery that the resulting typology of commercial patches reproduces a spatial hierarchical organization of consumption that evokes the classic Central Place Theory. Twelve major commercial corridors located in the central area of Singapore occupy the top of the hierarchy.