Research on the Spatial Distribution Characteristics and Influencing Factors of Educational Facilities Based on POI Data: A Case Study of the Guangdong–Hong Kong–Macao Greater Bay Area

Abstract

This study aims to provide a precise assessment of the distribution of educational facilities within the Guangdong–Hong Kong–Macao Greater Bay Area, serving as a crucial foundation for managing educational resource allocation and enhancing the quality of educational services. Utilizing a kernel density analysis, global autocorrelation analysis, and geographic detectors, this research systematically analyzes the spatial distribution characteristics and influencing factors of educational facilities in the area. The findings reveal significant geographical disparities in facility distribution with dense clusters in urban centers such as Guangzhou and Shenzhen, and less dense distributions in peripheral areas like Zhongshan and Macau. These facilities exhibit a multi-center cluster pattern with strong spatial autocorrelation, mainly influenced by the population size and economic and urban development levels. The results provide actionable insights for refining educational planning and resource allocation, contributing to the enhancement of educational quality across diverse urban landscapes.

1. Introduction

In the era of innovation-driven economic development powered by knowledge and technology, education is crucial for the potential and competitiveness of regional economies [1,2]. The allocation of resources for educational service facilities has become a key issue in urban planning and development management. Broadly defined, educational infrastructure includes all facilities necessary for the development of educational activities, encompassing both on-campus teaching resources and off-campus educational services [3,4]. As an essential component of educational services, these facilities significantly impact the development of educational endeavors [5,6]. Typically orchestrated and funded by the government, these infrastructures are designed to provide quality educational services to children, adolescents, and learners within communities [7,8]. Educational facilities serve the public, embodying public interest, universality, and extensive impact. The importance of educational services lies in their ability to benefit every individual and unit within society. Hence, educational infrastructure should be strategically designed to ensure the equitable distribution of educational resources and establish a fair public service system that upholds every city dweller’s right to education [9].

Current research on educational facilities generally focuses on two aspects: management and utilization. Regarding management, Nurhuda qualitatively analyzed educational facility management from the perspectives of planning, administration, and supervision [10]. Adiarti and colleagues explored how facilities impact student performance and the optimal operation of infrastructure [11]. Samanhudi argued that the planning of educational facilities should align with the vision and mission of educational institutions [12]. In terms of usage, Siljeg assessed satisfaction with the availability of educational facilities and services using descriptive statistics [13]. Hasbullah and colleagues developed a framework for assessing the performance of school facilities [14]. Setyono and colleagues measured the service rate and modes of urban educational facilities [15]. Huang and colleagues assessed the spatial equity of three types of basic educational facilities (kindergartens, elementary schools, and middle schools) [16]. Yenisetty applied spatial assessment techniques to study the accessibility of educational facilities [17]. Utomo used the neighborhood unit method combined with GIS technology to analyze the spatial distribution of community educational facilities in Palu city [18]. These studies typically focus on either management or usage aspects, rarely integrating both, which reveals a gap in comprehensive analyses. Moreover, the methods employed in past research have leaned heavily on literature reviews and theoretical exploration, with a lack of objective and specific quantitative analyses, particularly in the realm of facility management.

Due to the need to situate and construct educational facilities within a geographical space, these facilities inherently possess strong spatial attributes, containing vast amounts of geographic information. Previous research on the management of educational facilities has often overlooked this characteristic, which has contributed significantly to the difficulty in achieving a quantitative analysis in this area. Compared to other educational support services, the spatial configuration of educational facilities is a crucial aspect that affects the quality and level of educational services [19,20]. The spatial configuration of educational facilities is closely linked to a city’s overall development level and resource conditions. Due to differences in the agglomeration and radiation effects between cities, and variations in historical backgrounds and resource conditions, there might be imbalances in the development of educational facilities across different cities—as with other social resources. Focusing on the spatial distribution patterns of educational facilities is an important aspect of addressing real educational needs and uncovering potential imbalances in regional facility resource development. This study aims to integrate geographic information system (GIS) technologies into the field of educational facility management [21,22,23,24,25]. By analyzing and processing the spatial geographic information of educational facilities, it seeks to understand the spatial distribution patterns within specific areas and identify factors influencing these patterns, thus achieving quantitative analysis in educational facility management [26]. This provides precise data support for administrators to optimize facility resource allocation. Considering China’s vast urban expanses and populations, regional development issues are often studied within the context of urban clusters in a defined geographical area. This research focuses on the Guangdong–Hong Kong–Macao Greater Bay Area, one of the world’s four major urban clusters, known for its high level of economic development, dense urbanization, and rich educational resources. As a rapidly developing world-class urban cluster, the optimal configuration of educational facilities here can serve as a model for other regions. Therefore, using the Greater Bay Area as a case study, this research first employs GIS technology to visualize and analyze the current spatial distribution of educational facilities using Point of Interest (POI) data. It then uses geographic detectors to assess the impact of socioeconomic indicators on distribution patterns. Finally, it proposes optimization strategies for educational facility resource allocation. This study aims to provide scientific, targeted management insights for decision-makers and relevant departments, contributing to the high-quality, coordinated development of the Guangdong–Hong Kong–Macao Greater Bay Area by improving the current state of educational facility resource allocation [27,28].

This article starts by introducing the research background, highlighting the practical significance of studying educational facilities and outlining the current shortcomings and objectives of existing research. This is followed by a detailed description of the methodologies used in this study, including methods for identifying types of spatial distributions, analyzing density features, assessing correlation characteristics, and employing geographic detector tools from geographic information technology to evaluate influencing factors. Next, the Section 3 presents the findings of four subsections: the spatial distribution of types of educational facilities, their density characteristics, the assessment of spatial correlation features, and the main factors influencing the distribution of educational facilities. Based on these results, this study delves deeper into the practical implications of the revealed distribution patterns of educational facilities and proposes optimization strategies for the issues identified. This article closes with a summary of the main findings and contributions, emphasizing the practical value of using GIS technology in planning educational facilities, and suggesting directions for future research. Through the close connection and logical progression of its chapters, this study aims to provide scientific and targeted management bases for policymakers and relevant departments, enhancing the configuration of educational facility resources and facilitating high-quality, collaborative development in the Guangdong–Hong Kong–Macao Greater Bay Area.

2. Materials and Methods

2.1. Data Collection

The study area encompasses the 11 administrative regions of China’s Guangdong–Hong Kong–Macao Greater Bay Area, focusing on the educational facilities in the Greater Bay Area in 2022. Utilizing the Geocoding API platform provided by Gaode Maps and Baidu Maps data sources, a geocoding method was employed to input keywords and obtain information in bulk [29]. The search results include four attributes: name, address, coordinates, and category. After performing data cleaning, elimination, coordinate correction and address information correction, a total of 132,000 valid location records remained. The vector boundary data of each administrative region were obtained from the National Geospatial Information Public Service Platform (https://www.tianditu.gov.cn/, accessed on 9 June 2023). It should be noted that the educational facilities discussed in this paper encompass a broad range of facilities, including on-campus facilities such as kindergartens, primary schools, and secondary schools, as well as off-campus facilities with educational capabilities such as training institutions and libraries.

2.2. Statistical Methods

2.2.1. Imbalance Index

The imbalance index (S) is an important indicator that reflects the equilibrium state of the distribution of the research object within a region. The value of S ranges from 0 to 1, with higher values indicating a greater degree of imbalance in the distribution of educational facilities in the study area [30]. The calculation formula for S is as follows:

$$S=\frac{\sum _{i=1}^n Y_i-50(n+1)}{100n-50(n+1)}$$

2.2.2. Geographical Concentration Index

The geographic concentration index (G) is an important indicator that reflects the degree of geographical distribution concentration of the research object. The value of G ranges from 0 to 100. A higher value of G indicates a greater concentration of spatial distribution of educational facilities, while a lower value indicates a more dispersed distribution [31]. The calculation formula for G is as follows:

$$\mathrm{G}=100\times \sqrt{\sum _{\mathrm{i}=1}^{\mathrm{n}} (\frac{{\mathrm{x}}_{\mathrm{i}}}{\mathrm{T}}{)}^2}{|}_{\omega }$$

2.2.3. Nearest Neighbor Index

The nearest neighbor index is a geographic indicator that represents the degree of spatial adjacency in a geographical area. It is utilized to characterize the type of spatial distribution of educational facilities [32]. The calculation formula for the nearest neighbor index is as follows:

$$R=\overline{r}/r_j{r}_j=1/2\sqrt{n/A}=1/2\sqrt{D}$$

2.2.4. Nuclear Density Analysis

For point features, kernel density analysis is a spatial analysis method that reveals the spatial variation characteristics of point feature density. It calculates the density of point features around each output raster cell within its surrounding neighborhood, resulting in a smoothed raster surface that visually represents the aggregated form and density of the point feature distribution [33,34]. The calculation formula for kernel density analysis is as follows:

$$\mathrm{g}(\mathrm{x})=\frac{1}{\mathrm{n}\mathrm{h}}\sum _{\mathrm{i}=1}^{\mathrm{n}} \mathrm{k}(\frac{\mathrm{x}\cdot {\mathrm{x}}_{\mathrm{i}}}{\mathrm{h}}{)}^{\leftrightarrow }$$

2.2.5. Global Autocorrelation Analysis (Moran’s I Index)

Moran’s I index is an important indicator for identifying and measuring the spatial distribution relationships among units in a research area. When Moran’s I > 0, it indicates a positive spatial correlation among the units, whereas a negative correlation is indicated when Moran’s I < 0. Moreover, a larger value of Moran’s I suggests a more significant spatial correlation. This index is tested for significance using the Z-Score. If the Z-Score exceeds the critical value of 1.96, it indicates statistically significant spatial clustering, with a less than 5% chance of randomly generating such clustering patterns [35,36]. The calculation formula for Moran’s I index is as follows:

$$\mathrm{Moran}’\mathrm{s} I=\frac{n\sum _i \sum _j W_{ij}(X_i-\overline{X})(X_j-\overline{X})}{\sum _i \sum _j W_{ij}{\left(X_j-\overline{X}\right)}^2}$$

$$z_I=\frac{I+\frac{1}{n-1}}{\sqrt{V\left[I\right]}}$$

2.2.6. Local Spatial Autocorrelation Analysis

The application of the Getis-Ord G_i^* tool enables hot spot analysis, which visualizes the specific manifestation of the correlation between geographic phenomena or attribute values in a regional unit and those in neighboring units based on the spatial correlation measured by Moran’s I index [37,38]. The calculation formula is as follows:

$$G_i^{*}=\frac{\sum _{j=1}^n W_{ij}{X}_j-\overline{X}\sum _{j=1}^n W_{ij}}{S\sqrt{\frac{n\sum _{j=1}^n W_{ij}^2-{\left(\sum _{j=1}^n W_{ij}\right)}^2}{n-1}}}$$

2.2.7. Geographic Detectors

As an emerging statistical method, geographic detectors can address multicollinearity issues. They can calculate the influence of various potential factors on spatial heterogeneity and detect the interactions among these factors [39,40]. The calculation formula for geographic detectors is as follows:

$$\begin{array}{c}q=1-\frac{\sum _{h=1}^L N_h{\sigma }_h^2}{N{\sigma }^2}=1-\frac{SSW}{SST}\\ SSW={\displaystyle \sum _{h=1}^L} N_h{\sigma }_h^2,SST=N{\sigma }^2\end{array}$$

Among them, the q value is used to measure the extent to which factor X explains the spatial heterogeneity of Y. Additionally, due to the restrictions of variable data types in the calculation rules of geographic detectors, it is necessary to convert numerical variables into categorical variables using the K-means clustering method in IBM SPSS Statistics for Windows, version 27.0 (IBMCorp., Armonk, NY, USA) before processing the data.

2.3. Methods and Procedure

The research plan is structured into four subsections: spatial distribution types, density characteristics, correlation features, and analysis of influencing factors [41]. First, in the spatial distribution types section, this study will employ the imbalance index to measure the equity of educational facility distribution, reflecting whether resource allocation is fair. It will use the geographic concentration index to gauge the concentration level of educational facilities within a specific area, aiding in the preliminary identification of regions with dense facility distribution [42,43]. Additionally, the nearest neighbor index will be utilized by calculating the distance from each facility to its nearest neighbor, thereby determining the overall characteristics of the spatial distribution of educational facilities in the Guangdong–Hong Kong–Macao Greater Bay Area and describing their spatial distribution type. Following this, in the density characteristics section, this research applies kernel density analysis to transform a series of discrete point groups into continuous density data. By calculating the point density within a certain radius around each point, a smooth density map is generated. This method visually displays the spatial clustering characteristics. Next, in the correlation features section, the spatial proximity of the study subjects is considered by incorporating spatial autocorrelation analysis (Moran’s I index) to measure the overall spatial autocorrelation of the region [44]. A positive Moran’s I value indicates a clustering tendency among facilities, whereas a negative value suggests a dispersal trend. Local hot and cold spot analysis is then used to further analyze the interaction between units in local areas, identifying significant high-value (hotspot) and low-value (cold spot) areas, and visualizing the clustering locations of these areas [45]. Finally, in the analysis of influencing factors section, geographic detectors are used to analyze the specific impact of socioeconomic factors on spatial distribution. By measuring the explanatory power of different factors, geographic detectors quantify each factor’s contribution to the distribution of facilities, revealing the effects of variables such as population size and economic level on the distribution of educational resources. These methods complement each other and provide a comprehensive, detailed description of the spatial distribution characteristics of educational facilities in the Guangdong–Hong Kong–Macao Greater Bay Area and their influencing factors.

2.4. Selection of Influencing Factors

Based on the analysis of other relevant infrastructure impact mechanisms and considering the characteristics of educational facility development in the Greater Bay Area, this study selects nine social and economic indicators that may influence the development of educational facilities as the indicator system for exploring facility spatial distribution. Among them, the urbanization rate (X1) refers to the proportion of the population migrating to urban centers in a region, reflecting the extent of urban scale and expansion from a social attribute perspective [46,47,48]. The nighttime light index (X2) represents the brightness of surface lighting at night, essentially reflecting human activity intensity in terrestrial space, and is often considered an important indicator of urban development vitality [49,50]. Land area (X3) refers to the geographical space occupied by cities and serves as an indicator reflecting the scope of urban development and construction from a natural attribute perspective. Population indicators are often regarded as indicators determining the demand for educational facilities. Therefore, this study selects three related indicators: the number of urban residents (X4), population density (X5), and the number of students enrolled in schools (X6), aiming to comprehensively analyze the impact of population distribution on the distribution of educational facilities from different aspects [51]. Total road mileage (X7) is used as an indicator representing the level of urban development and construction. Regional gross domestic product (X8) and government education financial investment (X9) are selected as indicators measuring the level of economic development and investment in cities [52,53,54].

3. Results

3.1. Spatial Distribution Types

Comparing facility counts across different regions provides an initial understanding of the current configuration of educational facilities in the Greater Bay Area. According to the collected POI data, there is significant unevenness in the development status of educational facilities across different areas. Guangzhou (35,600) and Shenzhen (29,800) far exceed the other nine locations, holding a dominant position. Dongguan, Foshan, Huizhou, and Hong Kong range from 9000 to 20,000, showing relative advantages. Zhongshan, Jiangmen, Zhaoqing, Zhuhai, and Macau range from 700 to 7000, indicating the weakest level of infrastructure development and significant developmental disadvantages.

Using three spatial statistical methods—the imbalance index, geographical concentration index, and nearest neighbor index—under a premise of significant test validation, this study quantitatively describes the geographic distribution characteristics of educational facilities in the Greater Bay Area. The results show that the imbalance index (S) is 0.46, indicating significant unevenness; the geographical concentration index (G) is 38.34, suggesting high concentration; and the nearest neighbor index (NNI) is 0.21, with a Z-Score of −572.6, indicating a highly clustered distribution pattern within a few regions.

3.2. Density Characteristics of Spatial Distribution

After identifying the regional imbalances and localized clustering in the configuration of educational facilities, a kernel density analysis was employed using ArcGIS 10.2 software to further quantify this state. The natural breaks method was used to grade the density, visually representing the spatial clustering density of the educational facilities in the Greater Bay Area. As shown in Figure 1, the density distribution is relatively concentrated in the central–eastern parts of the region, exhibiting a non-symmetric polarization from the centers to peripheries. This diffusion follows a distance decay rule, resulting in three high-density and four medium-density areas, forming a “multi-center clustering” distribution state. The high-density areas center around Guangzhou, Shenzhen, and Hong Kong, while the medium-density areas center around Zhongshan, Huizhou, Dongguan, and the Zhuhai–Macau area.

Using a geographic detector, the analysis of the nine influencing factors (shown in

Table 1. Explanatory degree of impact factors on the distribution of educational facilities in the Greater Bay Area.

	X1	X2	X3	X4	X5	X6	X7	X8	X9
q statistic	0.11	0.30	0.37	0.96	0.37	0.87	0.16	0.65	0.77

) reveals that the explanatory power of these factors on the spatial distribution of educational facilities in the Greater Bay Area is ranked from highest to lowest as follows: number of permanent residents (X4) > number of students enrolled (X6) > government educational financial input (X9) > regional GDP (X8) > land area (X3) > population density (X5) > nighttime light index (X2) > total road mileage (X7) > urbanization rate (X1).

The q-values of these factors are compared in Table 1, showing a distinct hierarchy among the explanatory levels. The population-related indicators (X4, X5, X6) have the strongest explanatory power and are considered the dominant factors influencing the distribution of educational facilities. Economic indicators (X8, X9) show relatively strong explanatory levels and rank second in influence. Lastly, indicators reflecting urban development and construction levels (X1, X2, X3, X7) exhibit the weakest explanatory power, only modestly demonstrating their association with the distribution of educational facilities.

3.3. Spatial Correlation Characteristics

Using the fishnet tool in ArcGIS 10.2, uniform grid analysis units were constructed for spatial autocorrelation analysis, exploring the spatial relational characteristics of educational facilities in the Greater Bay Area. Moran’s I is 0.52, significant at p = 0.0000 and Z = 24.31, indicating strong global spatial autocorrelation and proximity effects among high-density areas.

Spatially, the Greater Bay Area educational facilities are divided into four clustering types: high–high, low–high, high–low, and low–low. High–high areas, showing strong developmental synergy, are primarily in central areas like Guangzhou, Foshan, Dongguan, Shenzhen, and surrounding Hong Kong. High–low areas, transitional zones with fewer distributions, are scattered and surrounded by extensive cold spots, indicating dependency and irreplaceability in certain regional areas.

The high–high type indicates that both the area and its surrounding regions are highly developed, exhibiting diffusion effects in spatial correlation, and demonstrating close regional interconnections that promote mutual development. The results map (Figure 2) shows that the high–high clusters are primarily located in the central areas of the Bay Area, forming a large-scale region around cities such as Guangzhou, Foshan, Dongguan, Shenzhen, and Hong Kong.

The high–low type represents areas where the local development level is high, but the surrounding areas are less developed. In spatial relationships, these areas act as transitional zones, capable of shifting towards either low–low or high–high types. As shown on the map in Figure 2, the high–low types are less frequent and are scattered, surrounded by extensive cold spots, showing significant polarization. These areas have strong dependencies on their surrounding cold spot regions and hold an irreplaceable dominant position within a certain geographic extent. The low–high type indicates regions with weaker development but surrounded by higher development levels. Figure 2 shows that the low–high distribution types are embedded around major cities like Guangzhou, Hong Kong, and Shenzhen. Being near high-value areas, these regions could potentially leverage the diffusion effects from the high–high clusters to achieve high-quality integrated development with adjacent areas. However, it is essential to avoid the ‘siphoning’ effect on resources such as population and capital flow from these high-value areas.

Conversely, the low–low type indicates that both the area and its surroundings have minimal infrastructure provision, representing the more disadvantaged units in the development of educational facilities in the Bay Area. As seen on the map, they are predominantly located on the periphery of the Bay Area and cover a large area. These vast urban outskirts have sparse distributions and insufficient supply of educational resources. In contrast, small urban core areas accumulate high-density infrastructure, leading to an oversupply. The pace of educational resource allocation lags behind the speed of urban expansion, resulting in imbalances within the internal subsystems of educational facility development.

3.4. Analysis of Influencing Factors

The application of geographic detectors to nine influencing factors reveals their varying degrees of explanatory power for the spatial distribution of educational facilities in the Greater Bay Area, as shown in Table 1. The factors are ranked from most to least influential as follows: number of permanent residents (X4) > number of students enrolled (X6) > government education funding (X9) > regional gross domestic product (GDP) (X8) > land area (X3) > population density (X5) > nighttime light index (X2) > total road mileage (X7) > urbanization rate (X1). A comparison of the q-values for these factors (Table 1) reveals a distinct hierarchy. Specifically, demographic indicators (X4, X5, X6) possess the strongest explanatory power and are considered the primary drivers of the spatial distribution of educational facilities. This is followed by economic indicators (X8, X9), which show a relatively strong level of explanation and rank second in terms of impact. Lastly, indicators representing urban development and construction levels (X1, X2, X3, X7) have the weakest explanatory power, demonstrating only a low level of association with the distribution of educational facilities.

4. Discussion

4.1. Discussion of Results

Following a thorough examination of the current state of the educational facilities in the Guangdong–Hong Kong–Macao Greater Bay Area, this study identified that the areas of high density for educational facilities are concentrated in the central and eastern regions of Guangzhou, Shenzhen, and Hong Kong. These regions have developed into a “multi-center clustering” distribution pattern. Cities like Guangzhou and Shenzhen, which are densely populated economic hubs that have been key focuses of China’s development efforts, often receive more policy support and investment in economic and educational resources due to their historical and resource contexts. These political and historical factors have laid a foundational basis for the current clustered distribution of educational facilities. The spatial autocorrelation analysis indicates that the high–high-type distribution areas are predominantly situated around Guangzhou, Foshan, Dongguan, Shenzhen, and Hong Kong, demonstrating strong regional connections and diffusion effects. Conversely, the low–low type areas are concentrated on the periphery of the Bay Area, where the facility configuration levels are weaker. This distribution highlights the advantages of the core cities in managing educational resources and their capacity to attract talent and economic investments, thereby facilitating more resource input for the construction of educational infrastructure [41,42,55].

The analysis of the influencing factors reveals that the number of permanent residents (X4) and the number of enrolled students (X6) have the highest impact on the spatial distribution, indicating that population size is the primary condition affecting the facility layout in the Bay Area. Because cities are a social public service product, the usage needs of the public must be considered in the layout of educational facilities [56]. Moreover, this indicates that the construction of educational facilities in the Bay Area generally matches the usage demands of its population size, suggesting that the spatial layout differences displayed by the educational facilities in the Bay Area are reasonably and scientifically justified. However, compared to the number of permanent residents (X4) and the number of enrolled students (X6), the population density (X5) did not show a high level of alignment. The explanatory power of the population density on the spatial distribution of the educational facilities in the Bay Area is relatively poor. The difference in the area size of the different cities shows that urban area size has disrupted the positive correlation of the population density indicator with the facility distribution. Economic factors (X8X9) provide the material basis for infrastructure construction. However, in terms of the explanatory power, economic factors do not have as direct and decisive an impact on facility demand as population factors (X4X6). Nevertheless, the results of detecting the interactive effects of influencing factors (

Table 2. Explanatory degree of interactive effects of influencing factors on the distribution of educational facilities in the Greater Bay Area.

	X1	X2	X3	X4	X5	X6	X7	X8	X9
X1	0.11
X2	0.38	0.30
X3	0.59	0.67	0.37
X4	0.99	0.98	0.98	0.96
X5	0.54	0.52	0.82	0.98	0.37
X6	0.92	0.91	0.91	0.97	0.98	0.87
X7	0.62	0.79	0.58	0.98	0.63	0.98	0.16
X8	0.99	0.98	0.97	0.98	0.98	0.98	0.98	0.65
X9	0.83	0.82	0.81	0.98	0.98	0.98	0.81	0.95	0.77

) show that economic factors (X8X9) have the best interactive explanatory effect on the distribution of educational facilities. When city development indicators such as the urbanization rate (X1), nighttime light index (X2), land area (X3), population density (X5), and total road mileage (X7) individually explain the facility distribution, their impact on the facility distribution is minimal. However, when these indicators are combined with economic factors for a joint effect analysis, the explanatory power of city development indicators (X1X2X3X5X7) on the facility distribution increases to above 0.8. This indicates that, although economic factors and city development levels are not as directly related as population size, they have a clear causal relationship with the distribution of educational facilities. This causality is best understood through joint analyses with economic factors and city development indicators to achieve the same explanatory effect as population indicators. The independent explanatory effect of government educational financial input (X9) surpasses that of the regional GDP (X8), yet the composite effect of the regional GDP with other factors outperforms the combination of educational spending with other factors, indicating a stronger direct relationship between educational financial spending and educational facility configuration [57,58,59,60].

4.2. Optimization Strategies

The findings of this study align with those of numerous existing studies that highlight that economically developed and densely populated regions often have more concentrated educational resources. Similar research has demonstrated that economic and demographic factors largely dictate the spatial distribution of educational facilities, a conclusion that is further substantiated by our analysis. Moreover, this study enhances our understanding of the spatial distribution of educational facilities and their influencing factors by incorporating kernel density analysis and geographic detector methods. The application of these methods not only validates existing research findings but also unveils more intricate spatial distribution patterns and mechanisms, offering new perspectives and tools for future research.

To clarify the practical implications and significance of the research findings, this study proposes four optimization strategies for future urban educational facility planning and management:

(1) Utilizing kernel density analysis and GIS technology allows for a visual display of the distribution density and spatial clustering characteristics of educational facilities. By developing a database for evaluating the spatial configuration levels of basic educational facilities that stores both spatial and non-spatial attribute data, GISs’ robust data processing and spatial analysis capabilities can be leveraged to analyze current issues and propose optimization strategies [61]. For instance, increasing the number of educational facilities in high-density areas and optimizing the utilization efficiency of existing facilities in low-density areas can enhance the overall service levels [62]. (2) Urban centers with high population densities may have schools that exceed standard sizes, while peripheral areas with low population densities struggle to meet their needs due to smaller educational scales. Therefore, planning standards should be adjusted according to the population density and growth in different areas to achieve rational resource distribution and effective utilization [28,63,64]. (3) By establishing digital cities and educational informatization platforms, the informatization of educational facility management can be achieved, enhancing government decision-making transparency and management efficiency. Online learning platforms and remote education technologies can partially address the inequalities in educational opportunities caused by geographical locations, especially in remote and underdeveloped areas [64,65]. By employing informatization, breaking down information barriers, and expanding the coverage of shared educational resources, the coordinated development of education between regions and urban–rural areas can be promoted [66]. (4) In newly developed residential areas, promoting a “large mixed-residence, small community” model that congregates residents from various economic strata within the same community can facilitate the sharing of educational resources [67]. This model is well suited for the renewal of old urban centers and helps to diminish residential space segregation, thereby improving the balanced allocation of educational facilities [68,69].

5. Conclusions and Contribution

5.1. Conclusions

This investigation into the spatial configuration of educational infrastructure within the Guangdong–Hong Kong–Macao Greater Bay Area has led to several distinct findings:

(1) There is a pronounced geographical pattern in the distribution of educational facilities, with denser concentrations in Guangzhou, Shenzhen, Dongguan, Foshan, Huizhou, and Hong Kong, compared to sparser distributions in Zhongshan, Jiangmen, Zhaoqing, Zhuhai, and Macau. (2) The facilities exhibit a clustered distribution, leading to regional polarization, primarily in the central and eastern parts of the area, showcasing a “multi-center cluster” with a notably asymmetric pattern. (3) The facilities’ spatial distribution is highly autocorrelated. Based on local distribution matching, the analysis identifies four distinct types of clusters, with hotspots including Guangzhou, Foshan, Dongguan, Shenzhen, and the surrounding areas of Hong Kong. (4) The population size primarily drives the spatial characteristics of educational facilities in the Bay Area, with secondary influences stemming from economic factors and levels of urban development.

This study interprets the spatial patterns of educational resources through a meticulous analysis of POI data and a detailed exploration of the spatial distribution of educational facilities. Analyses of socioeconomic indicators—such as the population size, economic development level, and urban comprehensive strength—deepen our understanding of the mechanisms that influence the distribution of educational resources, providing precise insights for the effective planning and allocation of educational facilities. These scientific insights serve as a valuable resource for policymakers, enhancing the coverage and quality of educational facilities, aiding in the resolution of spatial distribution issues, fulfilling diverse regional educational needs, and fostering educational equity.

5.2. Contribution

This study makes significant contributions in three key areas: First, the adoption of web crawling technologies for data collection overcomes traditional limitations associated with time and resource constraints, allowing for the extraction of a vast array of relevant, timely, and comprehensive data. The high-resolution and abundant sample size underpins a robust analysis of the status quo of educational facility distribution, enhancing the research’s comprehensiveness, accuracy, and credibility.

Second, the research results are innovatively presented using kernel density maps (Figure 1) and local spatial autocorrelation maps (Figure 2) that utilize color gradients and temperature changes to illustrate the facility distribution density and clustering intensity, respectively. This visual representation not only highlights the disparities in the educational resource distribution across different regions but also enhances the accessibility and clarity of the results through clear visual communication.

Lastly, the research employs advanced geospatial information statistical techniques to perform a comprehensive analysis from both the macro and micro perspectives, revealing the distribution patterns of educational facilities across this critical urban cluster. By moving beyond traditional single-factor analyses to include a multifaceted urban system framework, this study integrates external variables such as urbanization, economy, and population with the spatial distribution of educational facilities. This approach successfully elucidates the external environmental factors shaping the distribution of educational resources.

These findings provide valuable insights into the spatial distribution challenges of educational facilities, offering precise guidance for their strategic planning and allocation. This aids government and related management bodies in crafting development strategies that meet the needs of communities, thereby enhancing the management and quality of educational infrastructure.

6. Limitations and Further Prospects

This study employs various methods such as the imbalance index, kernel density analysis, and spatial autocorrelation analysis (Moran’s I index), which demonstrate significant advantages in revealing the distribution characteristics and influencing factors of educational facilities in the Guangdong–Hong Kong–Macao Greater Bay Area. However, there are certain limitations in the research [70,71]. Firstly, regarding the use of POI data, although precise geographic location data were collected, there are global issues concerning the quality of POI data that still exist, which may affect the accuracy of the analysis. Future research is needed to further verify the data quality and explore data cleaning and verification methods to enhance data reliability [21,23,25]. Additionally, this study primarily conducted an analysis based on static time points, failing to fully consider the impact of temporal changes on the distribution of educational facilities. Future research should expand into a broader time frame to explore the dynamic changes in educational facilities over time and to interpret the trends and patterns behind these changes. Moreover, the reliability of the kernel density analysis results is significantly affected by parameter selection, such as bandwidth, and so different parameter settings may lead to notable differences in results, necessitating parameter optimization in future research to ensure robustness. While spatial autocorrelation and local hot and cold spot analyses can reveal clusters of high and low values in specific areas, these methods are sensitive to outliers, which may compromise the robustness of the results. Furthermore, although the geographical detector effectively analyzes the impact of socioeconomic factors, it cannot address qualitative influencing factors such as policies, history, and culture, thus failing to fully reveal causal relationships [72]. Therefore, it is hoped that future studies will integrate more dynamic data, optimize parameter selection, and introduce more complex analytical methods to establish a more comprehensive and precise spatial optimization model, thereby enhancing the depth and breadth of the research [41,73,74].