How do people understand convenience-of-living in cities? A multiscale geographic investigation in Beijing

摘要

With the acceleration of global urbanization, especially for developing countries, more and more people live in cities, but their living environments are significantly different (e.g., slums and wealthy districts), resulting in massive social problems and attracting worldwide attention. This study aims at revealing the spatial heterogeneity of convenience-of-living (COL) in cities, and exploring how people consider COL from a geographic perspective. COL is defined based on the accessibility to diverse tangible amenities, e.g., parking lots, primary schools, and hospitals, and can be influenced by local built environments. To analyze COL, we first propose a segmentation method, i.e., fractal block net evolution, to spatially delineate analysis units for COL. This method can produce multiscale units adapting to different cognition scales of people. Then, we measure distances to diverse amenities as predictors and combine them with survey data by random forest regression, which can predict COL across Beijing city. As a result, a COL map is generated and clearly reports the spatial distribution of COL scores across this city which is highly heterogeneous but also has a large spatial autocorrelation. In addition, the experimental results indicate people that (1) mainly consider a neighboring scale while rating COL; (2) care more about bus/subway stations, restaurants, and shopping malls; and (3) follow some rules for scoring COL, which can be visualized by random forest. These findings contribute to explaining how people understand COL and assist infrastructure planning that can best satisfy the needs of local residents.