Fine-grained evaluation of neighborhood quality in China using street view images and big data technologies

With the advent of the post-materialist and post-industrialization era, quality of place has gradually become a hotspot of urban research. But the research has mainly focused on the city or national scale, and there is a lack of exploration on the small scale within the city. In this paper, we take neighborhoods as the research scale, design a neighborhood quality assessment system that can measure both the soft and hard environments of cities by using big data technologies such as deep learning and street view images to obtain microdata at the neighborhood level, and the entropy method was used to assess the quality of 5829 major neighborhoods (sub-districts) in 232 prefecture-level cities across China. The assessment results demonstrate that: (1) In China, the construction of "soft" environment is more important for enhancing the neighborhood quality. (2) Most of the high-quality neighborhoods in China are located in its economically developed regions, indicating that the enhancement of neighborhood quality is, to some extent, dependent on a region's economic level. (3) High-quality neighborhoods have been constructed in a relatively balanced manner in all dimensions, while low-quality neighborhoods show obvious deficiencies in the dimension of creativity. (4) Lastly, this paper includes a heterogeneity analysis on the neighborhood quality of cities in different regions, at different administrative levels, and with different industrial focuses, which reveals that even though China's construction efforts on "hard" environments of cities are pretty much the same, significant differences exist in those on "soft environment". This paper not only enriches the body of literature in the field of place quality but also provides significant theoretical support for urban planners and policymakers in planning and design, urban renewal, and the optimization of urbanization policies.