Abstract:

With the advancement of urbanization, reasonable allocation of public service resources is crucial for the sustainable development of cities and other regions. Education is the cornerstone of national revitalization and social progress; thus, the spatial distribution of educational resources should be in line with the overall trend of population distribution. This study attempts to evaluate the status of basic educational resource allocation on a finer scale based on population spatialization and POI (Point of Interest) data. We studied Changsha City in Hunan Province, where the 2018 census data were distributed into 100-m grids using POI, nighttime light, land use, road network, DEM (Digital Elevation Model), and other data as control variables based on the Random Forest model. We analyzed the spatial accessibility of basic education resources using cost weighted raster analysis and calculated the matching degree between the supply and demand of the resources using the improved potential model. The findings indicate that: 1) At the village (community) scale, the linear correlation coefficient (R²) of the population density between the result of population spatialization and the census data reached 0.73, which is better than WorldPop (World Population Dataset). Overall, the population distribution was displayed as a spatial distribution pattern of "one main-two sub-multi points", which could reflect the spatial characteristics of the population distribution in detail. 2) The spatial accessibility of the basic educational resources in Changsha was generally good: 97.72%, 96.20%, and 89.46% of the residents can reach the nearest primary schools, middle schools, and high schools, respectively, within 30 minutes. The difference in accessibility between the urban and rural areas was significant: the accessibility of the basic educational resources in facilities-intensive and traffic-developed areas was good and the villages (communities) with bad accessibility to basic education resources were concentrated in large counties with a large population of woodland and mountainous areas, especially in the surrounding mountainous areas of Liuyang County. 3) Generally, the basic educational resources in the urban areas of Changsha City and Liuyang and Ningxiang counties were in a balanced state of supply and demand; however, there were some areas where improvements were required. The matching degree between the supply and demand of basic education resources in towns was slightly higher, whereas in rural areas, it was mostly in a high-value state within the service radius and the intensity gradually decreased with facilities as the center. Therefore, it is necessary to strengthen the allocation of school buses and the ability of schools to receive boarding students to expand the scope of student sources. Previous studies mostly used large-scale census data, which ignored the spatial differences within each administrative unit. While some studies used small-scale census data such as those on residential quarters and buildings, however, these data are often difficult to obtain. This study provides research ideas for the precise assessment of the status of urban public service resource allocation and support for the diagnosis and improvement of the shortcomings of the basic educational resource allocation in Changsha City.

Key words: population spatialization, Random Forest model, POI data, basic educational resource, accessibility, matching degree, potential model, Changsha City