Speaker
Description
The concept of activity space is crucial in urban and behavioral geography, offering key insights for pedestrian-scale neighborhood planning, such as 15-minute neighborhoods or life-circles. However, traditional methods struggle to capture the complex and varied morphologies of activity spaces beyond simple circular forms. This paper presents a novel approach using topological data analysis (TDA), an advanced technique from applied mathematics and data science, to address these challenges. By integrating TDA with raster geocomputation, we redefine activity spaces as clusters and topological holes within a continuous field, reflecting the intensity of residents' daily activities. Thresholding strategies allow for precise and straightforward delineation. Utilizing mobile phone signaling data from Shenzhen, China, we employed TDA to uncover the boundaries of life-circles and explore their internal structures through both individual and community-level analyses, providing an accurate depiction of the morphology of activity spaces. We identified multiple detached components representing frequently visited places—anchor points or activity clusters, and cavities indicating areas of lower usage surrounded by frequently visited locations. The results demonstrated residents' spatial preferences for certain land use configurations and built environments, showing that life-circles have more diverse shapes and structures than simple circular forms. These findings can inform neighborhood planning practices by providing adaptive boundaries and configurations.
| Keywords | topological data analysis; activity space; life-circle; persistent homology |
|---|---|
| Best Congress Paper Award | No |
