POI-Based Assessment of Sustainable Commercial Development: Spatial Distribution Characteristics and Influencing Factors of Commercial Facilities Around Urumqi Metro Line 1 Stations

Abstract

Against the backdrop of rapid rail transit development, this study takes Urumqi Metro Line 1 as a case, using geographic information system (GIS) spatial analysis and space syntax Pearson correlation coefficient methods. Focusing on an 800 m radius around station areas, the research investigates the distribution characteristics of commercial facilities and the impact of metro development on commercial patterns through the quantitative analysis and distribution trends of points of interest (POI) data across different historical periods. The study reveals that following the opening of Urumqi Metro Line 1, commercial facilities have predominantly clustered around stations including Erdaoqiao, Nanmen, Beimen, Nanhu Square, Nanhu Beilu, Daxigou, and Sports Center, with kernel density values surging by 28–39%, indicating significantly enhanced commercial agglomeration. Metro construction has promoted commercial POI quantity growth and commercial sector enrichment. Surrounding commercial areas have developed rapidly after metro construction, with the most significant impacts observed in the catering, shopping, and residential-oriented living commercial sectors. After the construction of the subway, the distribution pattern of commercial facilities presents two kinds of aggregation patterns: one is the original centripetal aggregation layout before construction and further strengthened after construction; the other is the centripetal aggregation layout before construction and further weakened after construction, tending to the site level of face-like aggregation. The clustering characteristics of different business types vary. Factors such as subway accessibility, population density, and living infrastructure all impact the distribution of businesses around the subway. The impact of subway accessibility on commercial facilities varies by station infrastructure and urban area. The findings demonstrate how transit infrastructure development can catalyze sustainable urban form evolution by optimizing spatial resource allocation and fostering transportation–commerce synergy. It provides empirical support for applying the theory of transit-oriented development (TOD) in the urban planning of western developing regions. The research not only fills a research gap concerning the commercial space differentiation law of metro systems in megacities in arid areas but also provides a scientific decision-making basis for optimizing the spatial resource allocation of stations and realizing the synergistic development of transportation and commerce in the node cities along the “Belt and Road”.

1. Introduction

Rail transit, including metro (subway) and light rail transit, is an essential part of transportation systems in shaping modern metropolises. Cities like London, Paris, and Tokyo constructed their rail transit system alongside their metropolitan expansion to support the growth of land use, the development of industries, and population relief from the inner city. Due to the huge financial investment needed for its construction and the Chinese trajectory of modernization and urbanization, most regional central cities in China started to construct rail transit around 2010, except Beijing, Shanghai, Tianjin, and Guangzhou, which built their inner-city rail transit decades earlier. This is very different from the developed countries, for the rail transit of China is essentially being built after each metropolis’s expansion. From the experiences of Beijing and Shanghai, the development of rail transit not only relieves traffic congestion but also profoundly influences the distribution patterns of public facilities around the metro stations.

Meanwhile, because the history of rail transit development is short, most research is focused on Beijing and Shanghai, and little is known about inner and western cities. To fill in these research gaps, this paper takes Urumqi, the westernmost provincial capital with the rail transit system in China, as a sample to study the influence of rail transit on the distribution of commercial facilities around metro stations.

As the capital of Xinjiang Uygur Autonomous Region, China, Urumqi is the most densely populated city in the northwest boundary region of China, with 4.08 million people at the end of 2022. The rail transit system of Urumqi, which was planned in 2012, has four metro lines. Among the four lines, Urumqi Metro Line 1 started to be built in 2014, was completed in 2018, and is the only operational metro line in the city now. As Figure 1 indicates, Metro Line 1 runs from north to west across the most populated areas of Urumqi. Urumqi Line 1, as a rapid transportation line vertically connecting the comprehensive center in the south of the city and the sub-center in the north of the city, passes through the International Aviation Center (International Airport Station), the Sports and Leisure Center (Sports Center Station), the traditional commercial center (Xiaoximen Station), and the core area of the old city (Erdaoqiao Station) along the line, which has good infrastructure conditions and population base, and the construction of the subway line has provided not only convenient transportation conditions for the rapid connection of each center but also convenient transportation conditions for commercial businesses and the gathering of population. By connecting these centers, Metro Line 1 not only reduces the distance between people and facilities, and alleviates the traffic congestion in Urumqi, but also increases the competitiveness and attractiveness of the public facilities in these central areas.

This study takes Urumqi Metro Line 1 as a case study and uses historical POI data to quantitatively explore commercial facilities’ distribution characteristics and influencing factors around metro stations. The research employs kernel density analysis, standard deviation ellipse analysis, space syntax, and correlation analysis. Kernel density analysis visualizes the hot and cold spots of commercial facilities within the influence of metro stations, exploring the overall distribution pattern of commercial facilities; standard deviation ellipse analysis visualizes the relationship between the development direction of commercial facilities and metro lines before and after the construction of the metro lines; and spatial syntax further explores the role of accessibility in the distribution of commercial facilities through the computation of the local integration degree of the stations. Correlation analysis is conducted to explore the relationships between commercial facility distribution and living infrastructure (including the availability of public facilities, scenic spots, healthcare, government institutions, scientific, educational, cultural facilities, etc.).

Moreover, it seeks to determine how the analysis of commercial facility distribution and its influencing factors can guide public transportation planning and sustainable commercial development in Urumqi. This study fills gaps in the study of commercial distribution around metro stations in developing cities in western China by providing an in-depth understanding of the distribution characteristics of commercial facilities around Urumqi metro stations while at the same time providing practical reference suggestions for urban policymakers and planners. By further investigating the relationship between factors such as accessibility, population density, living infrastructure, and the distribution of commercial facilities, the study demonstrates the feasibility of previous findings in developing cities in western China. Thus, it can help to improve the layout of commercial implementation around Urumqi metro stations, the efficiency of public transportation, and the sustainable development of commerce.

2. Literature Review

2.1. Research on the Relationship Between Rail Transit and Urban Facilities

Currently, research on the relationship between rail transit and the distribution of urban facilities in China mainly focuses on the following aspects. First, many studies analyze the impact of rail transit on the distribution of commercial facilities by examining the agglomeration and spatial restructuring of facilities around metro stations and changes in the business formats of these facilities [1,2]. Chen et al. (2022) investigated the differences in transport accessibility changes between different communities under the influence of rail transport, assessed its spatial equity effect, and then provided theoretical support for the integration of multimodal urban public transport systems [3]. Liang, J. (2022) found that commercial facility density around rail transit stations is significantly higher than in other areas, and often forms commercial agglomeration zones [4]. Zhuang et al. (2022), through their study of Shanghai’s rail transit system, found a notable increase in the proportion of commercial land use around metro stations [5]. Furthermore, according to the Report on Commercial Development around Railway Stations in Shanghai (2019), there is a gradient in business formats around stations, shifting from traditional retail to high-end consumption and service industries. Second, the feedback effect of commercial facilities on rail transit. These studies primarily focus on issues such as the attraction of public transport flows and the functional diversification of stations. For example, Zhang et al. (2020) found that the spatial clustering of commercial facilities has an efficiency-enhancing effect on rail transit, especially in rail transit sub-centers or hubs [6]. Wang Li et al. (2023) [7] found that diversified commercial facilities can effectively increase the vitality and attractiveness of metro station areas, and such an effect is more significant under the transit-oriented development (TOD) model [8,9].

Furthermore, various studies have tried to find the factors or processes that may play a crucial role in shaping commercial distribution patterns around metro stations. (1) Some studies have shown that areas with good accessibility to metro stations are more likely to attract commercial investment and development, leading to higher facility density and diversity of commercial facilities [10]. For example, Cervero (2007) found that areas within a half-mile radius of metro stations in the San Francisco Bay Area had significantly higher commercial density compared to areas further away [11]. (2) Demographic characteristics and socio-economic factors also significantly influence commercial distribution. Of these, population density, income levels, and consumer behaviors are often considered as key factors influencing the combinational and distributional patterns of multiple commercial facilities. (3) Urban planning policies and regulations are crucial in shaping commercial distribution around metro stations. Policies promoting mixed-use development and encouraging commercial facilities development in transit-oriented areas can promote the vitality and attractiveness of station areas [12,13]. For example, mixed land-use regulations of the TOD model around subway stations in cities such as Tokyo and Hong Kong increase the competitiveness and attractiveness of urban centers in the station areas [14,15].

The studies above provide great insights into the links between rail transit and urban development. However, several omissions and issues also need to be noted. The most obvious issue is that nearly all the research takes the mature rail systems in metropolises along the eastern seaboard as its objects but neglects the developing cities in the western region. And quite a lot of studies are cross-sectional, lacking the necessary longitudinal analysis to identify and confirm the logic of those development relationships.

2.2. Utilization of POI Data in Urban Studies

The two significant attributes of points of interest (POIs) data make it quite popular and vital to the practice of geographic information systems (GIS) and spatial analysis across various disciplines. First, it has detailed location information, and second, it has rich attribute information about the POIs. By utilizing the information of POI data, researchers and practitioners can dig into the pattern of spatial phenomena, test theories and hypotheses, and discover spatial processes at multiple scales and levels. With the exponential growth of GPS sensors, ranging from mobile phones to semi-autonomous vehicles, points of interest (POIs) data are now easy to access and widely used in research and practice.

Given such exponential growth of GPS sensors, POI data have not only become readily available, but also has a wealth of information about businesses, landmarks, and public facilities that make it an important data support for the study of the distribution of urban commerce, which is widely used in related research and practice [16]. In China, the application of POI data has been widely recognized in various fields such as urban planning, socio-economic studies, environmental ecology, etc. In urban planning, POI data have been used to evaluate the accessibility of facility services [17,18], delineate urban functional zones [19], identify urban centers [20], and analyze the distribution characteristics of different industries. For example, based on POI data, one researcher analyzed the shortage of green space service space in Tacheng, using an improved Gaussian two-step moving search method, and then made recommendations for planning improvements [21]. In socio-economic research, by modifying the SEIR model of SARS, some scholars established a multivariate data model of epidemic transmission in Beijing, and found that the population, socio-economic factors, and medical capacity of Beijing are highly heterogeneous spatially, so as to assess the spatio-temporal risk level of each district in Beijing [22]. In environmental ecology, by using POI data, scholars have quantitatively investigated the impact of urban development on natural habitats of wild animals and plants, clarifying the extent of natural habitats and ecological reserves that need urgent protection [23]. In the field of public health, scholars have assessed the spatial accessibility of medical facilities through massive POI data, providing planning references for the spatial distribution of medical facilities [24,25].

With the development of artificial intelligence and machine learning, there are a lot of new algorithms and methods that can be applied in POI data mining now. For example, Chenxi Jin et al. used a machine learning classification algorithm to process POI data from three major urban agglomerations in China and grouped the manufacturing sectors into seven types [26]. Yet another research study explored the spatial distribution patterns of different types of POIs through spatial autocorrelation analysis, which revealed the potential spatial structure of the urban area [27]. The authors of [28,29] used functional structure-based AI techniques to analyze commercial AOI data, POI data, nighttime lighting data, and population distribution data to classify urban land use and provide suggestions for the efficient use of spatial resources.

In addition, integrating POI data with other types of big data (e.g., social media data and cell phone signal data) in research has great potentiality to improve and expand our understanding of urban dynamics. Future research should explore this method further.

3. Research Design and Methodology

Based on the POI data of commercial facilities in Urumqi in 2017 and 2022 (grouped into six categories, namely catering services, shopping services, financial services, living services, sports and leisure, and accommodation services), the article takes Tianshan District, Shaybak District, and New Urban District of Urumqi, where the metro is situated, as the large study area. It adopts the methods of the ArcGIS platform, such as kernel density analysis, standard deviation ellipse, and buffer zone analysis, to quantitatively evaluate the spatial distribution characteristics of the commercial facilities around the metro stations based on the spatial and temporal changes in the number of facilities; the methods of superposition analysis, correlation analysis, and spatial syntax are also used to further demonstrate the influence of the relationship between the commercial facilities and the population, the accessibility of the stations, and the other infrastructure. In order to determine the influence range of the metro station more reasonably, the article scientifically selects 800 m as the influence range of the metro station based on previous studies, Peter Calthorpe’s theory of transit-oriented development, and the Guidelines for Planning and Designing of Areas Along Urban Railways, and carries out the relevant statistical analyses within this influence range.

3.1. Scope of Research

As the capital city of Xinjiang Uygur Autonomous Region (XUAR), Urumqi has a total administrative area of 13,800 square kilometers, with seven municipal districts and one county under its jurisdiction. At the end of 2023, the city’s resident population was 4,084,800 people, and the region’s gross domestic product (GDP) reached 333,732 billion yuan (45.65 billion US dollars). Urumqi Metro Line 1, the first metro line built and operated in Urumqi City, opened the northern section of the project (International Airport Station to Balou Station) for operation on 25 October 2018, and opened the southern section of the project (Balou Station to Santunbei Station) for operation on 28 June 2019. Currently, it is the only subway line in Urumqi City to have opened, with a total mileage of 27.615 km. This paper takes Urumqi Metro Line 1 as the research object. Urumqi Metro Line 1 starts from Santunbei Station and runs to International Airport Station, passing through four administrative districts of Urumqi, with a total of 21 metro stations, of which three stations are interchange stations, as shown in Figure 2. Urumqi Metro Line 1 entered the scope of the city’s master plan in 2012, started construction in 2014, officially opened at the end of 2018, and is now fully operational.

Several existing methods and theories are commonly employed in determining the study area along the metro line. According to the transit-oriented development (TOD) theory proposed by Peter Calthorpe, the influence range of public transportation stations typically spans a walking distance of 5 to 15 min, corresponding to a radial distance of approximately 400 to 800 m. The Urban Rail Transit Area Planning and Design Guidelines classify the vicinity of rail transit stations into two categories: a core influence zone within 300 to 500 m and an extended influence zone within 500 to 800 m. Most related studies adopt an 800 m radius as the standard influence range for metro stations [30,31,32]. Based on the aforementioned theories, guidelines, and existing research, this study defines an 800 m radius around each metro station as the influence range for analysis.

3.2. Research Methodology

3.2.1. Kernel Density Analysis

In ArcGIS, kernel density analysis is a widely used spatial analysis tool [33,34]. The processed SHP data were extracted and analyzed within the defined study area to obtain POI data for six categories of life service industries along Urumqi Metro Line 1. Kernel density analysis is based on point features and explores spatial patterns using a fixed search radius. It calculates the density of points within a specified neighborhood, generating a continuous surface of density values. Different kernel density values reflect the spatial distribution characteristics of the target features. This method relies on the density function of the data and employs clustering algorithms to achieve spatial density analysis [35]. The formula for kernel density estimation is as follows:

$$\mathrm{f}\left(s\right)=\sum _{\mathrm{i}=1}^n{\displaystyle \frac{1}{{\mathrm{h}}^2}}\mathrm{k}\left({\displaystyle \frac{{\mathrm{d}}_{\mathrm{s}\mathrm{i}}}{\mathrm{h}}}\right)$$

(1)

where s represents a point feature; f(s) denotes the kernel density value of s; n indicates the number of features within the distance h from the target s; h represents the search radius; k signifies the spatial distance weight; and dsi represents the distance between POI point i and s.

3.2.2. Standard Deviation Elliptic (SDE) Analysis

The standard deviational ellipse is one of the spatial statistical techniques used to measure the distribution patterns of geographic features. It simultaneously analyzes the direction and distribution of points, quantitatively describing the spatiotemporal distribution characteristics of the study object through parameters such as the centroid, rotation angle θ, and standard deviations of the X and Y axes [36,37]. In ArcGIS, the Directional Distribution tool is used to create standard deviational ellipses to summarize the spatial characteristics of geographic features. Before analysis, the feature geometry is projected onto the output coordinate system. The output ellipse’s attribute values include two standard distances (major axis and minor axis) and the orientation of the ellipse, ultimately revealing the central tendency, dispersion, and directional trends. This analysis method is employed to explore the spatial distribution range and concentration degree of urban residents’ basic life service industries. The formula for the midpoint of an ellipse is shown in

Table 1. Formula for the midpoint of an ellipse.

No.	$\mathbf{Formula}$	Parameter Meaning	Applications in This Study
1	$SD{E}_x={\displaystyle \frac{\sum _{i=1}^n{\left(x_i-\overline{X}\right)}^2}{n}}$ $SD{E}_y={\displaystyle \frac{\sum _{i=1}^n{\left(y_i-\overline{Y}\right)}^2}{n}}$	In the formula, (SDEx, SDEy) are the coordinates of the centre of the ellipse, xi and yi are the coordinates of the spatial position of each element, (xi,yi) are the coordinates of the life service industry i of the ellipse, and X and Y are the centres of arithmetic mean.	Calculate the midpoint of the ellipse
2	$\mathrm{t}\mathrm{a}\mathrm{n}\theta ={\displaystyle \frac{A+B}{C}}$ $={\displaystyle \frac{\left(\sum _{i=1}^n\widetilde{x_i^2}-\sum _{i=1}^n\widetilde{y_i^2}\right)+\sqrt{{\left(\sum _{i=1}^n\widetilde{x_i^2}-\sum _{i=1}^n\widetilde{y_i^2}\right)}^2+4\left(\sum _{i=1}^n\widetilde{x_i}\widetilde{y_i}\right)}}{2\sum _{i=1}^n\widetilde{x_i}\widetilde{y_i}}}$	In the formula, θ is the azimuth angle, xi and yi are the difference between the mean centre of the ellipse and the xy coordinates.	Calculate azimuth
3	${\sigma }_x=\sqrt{2}{\displaystyle \frac{\sum _{i=1}^n{\left(\widetilde{x_i}\mathrm{c}\mathrm{o}\mathrm{s}\theta -\widetilde{y_i}\mathrm{s}\mathrm{i}\mathrm{n}\theta \right)}^2}{n}}$ ${\sigma }_y=\sqrt{2}{\displaystyle \frac{\sum _{i=1}^n{\left(\widetilde{x_i}\mathrm{s}\mathrm{i}\mathrm{n}\theta +\widetilde{y_i}\mathrm{c}\mathrm{o}\mathrm{s}\theta \right)}^2}{n}}$	In the formula, ${\mathsf{\sigma }}_{\mathrm{x}}$, ${\mathsf{\sigma }}_{\mathrm{y}}$ denotes the standard deviation along the long and short axes of the ellipse, respectively.	${\mathsf{\sigma }}_{\mathrm{x}}$ Calculate the value of the long half-axis of the ellipse, ${\sigma }_y$ Calculate the value of the short half-axis of the ellipse

.

3.2.3. Correlation Analysis

Correlation analysis is used to investigate the relationship between quantitative data, whether there is a relationship, and how strong the relationship is. The strength of a linear correlation can be measured by the magnitude of the Pearson correlation coefficient [38].

The correlation coefficient can be categorized as positive, negative, or having no correlation. If the Pearson correlation coefficient is between 1 and −1, the larger the absolute value of the coefficient, the greater the degree of correlation. |r| close to 1 means that the two have a perfect correlation; |r| close to 0 means that the two do not have a linear correlation (or possibly a non-linear correlation). In general, the spatial correlation coefficient r ranges from 0.8–1.0, 0.6–0.8, 0.4–0.6, 0.2–0.4, and 0–0.2, which correspond to very high, high, medium, low, and virtually no correlation between two features, respectively [39,40]. In this study, the Pearson correlation coefficient formula that comes with the GIS platform is used to realize the correlation relationship between the changes in the number of commercial facilities and the changes in the number of various types of amenity infrastructures. The Pearson correlation coefficient r is calculated as follows:

$$r={\displaystyle \frac{\sum _{i=1}^n\left(x_i-\overline{x}\right)\left(y_i-\overline{y}\right)}{\sqrt{\sum _{i=1}^n{\left(x_i-\overline{x}\right)}^2}\sqrt{\sum _{i=1}^n{\left(y_i-\overline{y}\right)}^2}}}$$

(2)

3.2.4. Accessibility Analysis

In studies related to urban facility distribution, accessibility is often considered a critical factor influencing the spatial distribution of urban facilities [41]. Accessibility refers to the ease with which residents can reach public facilities within a given time and cost range. It encompasses three dimensions: temporal accessibility, cost accessibility, and spatial accessibility. In urban planning and design, spatial accessibility is frequently calculated using space syntax [42].

Integration, a key analytical dimension in space syntax analysis, measures the degree of aggregation or dispersion between an element and other elements in a spatial system. It quantifies the ability of a space to attract traffic as a destination, reflecting its centrality within the overall system. Integration can be categorized into global integration and local integration. Spaces with higher integration values exhibit greater accessibility, stronger centrality, and a higher potential to attract pedestrian flow. The formula for calculating integration is as follows:

$${ I}_i={\displaystyle \frac{m\left[{\log }_2\left(\frac{m+2}{3}-1\right)+1\right]}{(\mathrm{m}-1)\left|\overline{D}-1\right|}}$$

(3)

where I_i represents integration; $\overline{\mathrm{D}}$ denotes the mean depth value; and m signifies the number of connections from a given space to other spaces.

3.3. Research Data

3.3.1. Data Processing

First, vector data for Urumqi Metro Line 1 (including rail alignments and stations) were extracted from the metro network dataset. Subsequently, POI data for six categories of commercial facilities were imported into a geographic information system (ArcGIS 10.4). Databases documenting spatial changes in these six commercial categories (2017 vs. 2022) were then constructed by delineating an 800 m radius buffer around metro stations as their spatial influence zone. Based on these databases, kernel density estimation and standard deviational ellipse analyses were conducted to examine temporal shifts and distribution trends in commercial agglomerations before and after metro construction. Finally, the integrated database was combined with supplementary datasets, including 2022 population statistics, urban road networks, and five categories of living infrastructure POI data within station influence zones. Through overlay analysis, spatial syntax modeling, and correlation analysis, this synthesis enabled further exploration of factors influencing commercial facility distribution around metro stations.

3.3.2. Data Sources

There are three main types of data used in this study; firstly, the POI data of Urumqi City in 2017 and 2022, obtained from “https://www.poi86.com/ (accessed on 7 January 2025)”, which includes six categories of data about business, such as catering services, shopping services, living services, financial services, sports and leisure services, and accommodation services, as shown in

Table 2. Six types of commercial facilities and the types of facilities they contain.

No.	Categories	Facility Types Included
1	Catering	Restaurants, pubs, etc.
2	Shopping services	Supermarkets, malls, shopping centers, etc.
3	Financial services	Banks, insurance companies, securities companies, etc.
4	Living services	Laundromats, barbershops, post offices (postal outlet), etc.
5	Sports and leisure services	KTV, bars, cinemas, fitness centers, etc.
6	Accommodation services	Hotels, five-star hotels, guest houses, etc.

; secondly, Urumqi City vector data, among which the population data were mainly from “https://www.iteye.com/ (accessed on 25 January 2025)” and the vector data of metro stations and lines were from “https://wenku.csdn.net/ (accessed on 25 January 2025)”; and thirdly, the base map data used in the study, incorporating the vector base map of Urumqi from “https://www.poi86.com/ (accessed on 7 January 2025)” and the underground line map of Urumqi from “https://www.urumqimtr.com/ (accessed on 15 January 2025)”. The overall research framework of this study is shown in Figure 3.

4. Results

4.1. Characteristics of Spatial Distribution of Commercial Facilities

(1)

Static distribution of commercial facilities

The static distribution of commercial facilities shows that the six types of commercial facilities in Urumqi are more concentrated in the southern section of the metro line (the core area of the old city) and relatively fewer in the northern section of the metro line (the station in the new city); among them, the facilities for catering, shopping, financial, and living services are mainly concentrated in the core area of the old city, and the facilities for sports and recreation and accommodation have a certain degree of distribution in both the old city and the new city. Different commercial facilities show different distribution characteristics according to the different basic conditions of the areas where they are located; in particular, the subway stations in the old city with perfect infrastructure and unique regional characteristics and cultural conditions show significant commercial aggregation trends.

(2)

Changes in kernel density of commercial facilities

The paper utilized historical POI data of Urumqi City to quantitatively explore the characteristics of the commercial spatial layout around the subway stations. Based on the kernel density analysis method in the ArcGIS platform, it carried out a pairwise comparative analysis of the changes in the spatial distribution of the kernel density of the commercial facilities within the 800 m influence area of the stations of subway line 1. The results are shown in Figure 4.

The comparative analysis found that after the completion of the subway, the nuclear density value of each site increased significantly, and the nuclear density gathering area is mainly concentrated in the southern part of the central city. Compared with before the completion of the subway, the overall nuclear density value of the station area has increased by 28–39%. However, it is worth noting that the kernel densities of different commercial facilities show different distribution and development trends in different station spaces, including the trend of centripetal aggregation distribution and the weakening of centripetal aggregation characteristics.

• Catering and sports and leisure facilities show obvious spatial aggregation characteristics in some station spaces. Through the horizontal comparison of the kernel density of various commercial facilities in Figure 3, it is found that the kernel density of catering facilities increased from 46.7 to 66.2 after the completion of the subway, and the aggregation characteristics are more obvious at South Gate Station and Xinxing Street Station. Similarly, the kernel density of sports and leisure facilities increased from 42.8 to 72.3, and their clustering characteristics are especially obvious at the stations of South Gate, North Gate, and South Lake Square on Xinxing Street. The average proximity value of the above two facilities to subway stations is less than 1, and the aggregation and the centripetal aggregation distribution characteristic is obvious.
• The centripetal aggregation of shopping, living, accommodation, and financial facilities in the overall site space has weakened and tends to be increasingly spread around the site level. The kernel density of shopping facilities decreased from 67.2 to 58.3; the kernel density of living facilities decreased from 47.8 to 42.1; the kernel density of accommodation facilities decreased from 57.4 to 52.6; and the kernel density of financial facilities decreased from 42.3 to 38.6.

In summary, the construction of the subway line has not only promoted the further spatial aggregation of some commercial facilities but also gradually weakened their original aggregation characteristics.

4.2. Changes in the Number of Commercial Facilities

To further explore the changes in commercial space before and after the construction of the subway, the historical POI data of commercial facilities were extracted and spatially visualized for the top eight subway stations with the top eight kernel density values of commercial facilities. The results are shown in Figure 5 within the 200 m, 500 m, 600 m, and 800 m influence ranges of the subway stations.

(1)

Changes in the number of commercial facility POIs in the 800 m influence area

With the construction of the subway lines, the number of facilities within the sphere of influence of the major subway stations increased significantly, with the number of dining facilities, shopping services, and lodging facilities increasing more significantly by 45%, 37.3%, and 58.8%, respectively (as shown in Figure 5a). In other words, the construction of the subway has contributed to a further increase in the number of dining, shopping, and accommodation facilities.

(2)

POI changes at the site level

With the construction of the subway, the number of POIs of commercial facilities such as catering, shopping, and living services at the South Gate, North Gate, and Xinxing Street stations has increased significantly. The commercial facilities around these stations have gradually tended to diffuse on the surface from the original point-like or band-like aggregation. The spatial scope of the commercial layout has gradually expanded (as shown in Figure 5b). Commercial facilities show a differentiated distribution pattern within different influence ranges (200 m, 500 m, 600 m, 800 m) of the subway stations. Overall, the development of the subway has contributed positively to the growth and diversification of commercial activities.

Figure 5. (a) Statistics on changes in the number of POIs of commercial facilities within the 800 m influence area of subway station sites. (b) Spatial distribution of commercial facilities within different spheres of influence of metro stations.

Figure 5. (a) Statistics on changes in the number of POIs of commercial facilities within the 800 m influence area of subway station sites. (b) Spatial distribution of commercial facilities within different spheres of influence of metro stations.

4.3. Direction of Development of the Distribution of Commercial Facilities

In order to further clarify the development direction of the spatial distribution of the surrounding commercial facilities before and after the construction of the subway, this article uses historical POI data within the 800 m influence range of the main subway stations and the standard deviation ellipse analysis method in ArcGIS to calculate the standard deviation ellipses before and after the construction of the subway in each of the main stations. The results of the calculations are shown in Figure 6.

In the figure, the major axis of the SDEs represents the primary developmental orientation of geographic elements, with a mean angular deviation of 10.3° (p < 0.05). The results reveal significant changes in the SDEs before and after the metro line construction. Compared with 2017, the major axes of the SDEs for stations such as Erdaoqiao, Nanmen, Xiaoxigou, Tieluju, and Sangong in 2022 exhibit notable shifts toward the direction of the metro line. Meanwhile, 78% of new commercial POIs occur within the 800-metre station buffer zone along the rail corridor. Most of the above major metro stations are located in Urumqi’s traditional commercial, tourism, and urban sub-centers areas, where the number and completeness of the original infrastructure are significantly better than in other parts of the city. Under the above basic conditions, the construction of the metro line has brought more convenient transportation conditions to the areas along the metro line, which further improves the commercial location advantage of the region, thereby attracting more commercial businesses to gather along the line.

The results show that since the completion of the subway, there has been a certain consistency between the spatial development direction of commercial facilities and the direction of the subway line, indicating that the construction of the subway line has produced a spatial agglomeration effect on the distribution of commercial facilities along the line.

4.4. Influencing Factors

4.4.1. Population Density Factor

In related studies, the population factor is considered to be the main factor affecting the spatial distribution of commercial facilities [35,43]; in order to further verify the correlation between the spatial distribution of commercial facilities around the Urumqi metro stations and the population factor, the spatial element superposition analysis method in ArcGIS was used to superimpose the 2022 population vector data of Urumqi with the influence range of the metro station. The results of the analysis are shown in Figure 7.

Figure 7 shows that the spatial distribution of high-density population areas is highly consistent with the direction of the subway lines. Also, the population density within the 800 m influence area of the metro stations (5600 persons/km²) is higher than the average population density of the study area (3000 persons/km²). Significant spatial overlap exists between the population hotspots and the subway stations.

In other words, the rapid development of the subway has increased commercial activity, promoted the penetration of pedestrian flow into the commercial area, and is driving the distribution of commercial agglomeration.

4.4.2. Living Infrastructure Factor

To further explore the relationship between the spatial distribution of commercial facilities around metro stations and the influencing factors, the Pearson correlation coefficient method in ArcGIS was used to establish the correlation coefficient matrix of the changes in the number of the six types of commercial facilities and the living infrastructure within the influence of the metro stations (2017 vs. 2022), and the results are shown in Figure 8. The significance level in the figure is less than 0.05, which means that there is a significant correlation between the variables.

As shown in Figure 8, the correlation coefficients between changes in the number of commercial facilities and changes in the number of public facilities and scenic spots are 0.791 and 0.71, respectively, with both correlation test values being less than 0.05. It shows that there is a positive correlation between the changes in the number of commercial facilities and the changes in some infrastructures, such as public facilities, scenic spots, and educational, cultural, and sports facilities; among them, the changes in commercial facilities are significantly positively correlated with the changes in public facilities and scenic spots, and negatively correlated with the medical facilities and administrative institutions.

Urumqi is not only the first populous city in the development of Xinjiang but also a national tourist destination; in this context, the good development of the city can not be separated from the improvement in the conditions of public facilities and the foundation of tourism facilities. At the same time, there is a specific incompatibility between commercial facilities and parts of the living infrastructure in terms of function, such as hospitals, CDC, and other areas where there is a risk of disease infection or mental depression; administrative agencies require a more serious and quiet office environment, which is incompatible with the bustle of the commercial environment.

4.4.3. Accessibility Factor

In the following, from the perspective of spatial accessibility, based on the local integration degree value in the spatial syntax analysis, the influence relationship between the accessibility of subway stations and the spatial distribution of the surrounding commercial facilities is further demonstrated. After cleaning and single-line processing of the road network vector data of the study area, it was imported into depth map software. The local integration degree analysis was carried out with a station influence range of 800 m.

As shown in Figure 9, most subway stations are in an area with a high degree of road integration. It was found that most of the site spaces with high accessibility are located at the core of the urban road network (road network intersections, around major transportation stations, etc.), with convenient transportation and directly connected to other spaces. Such areas are more likely to attract people flows, with more potential customers, and are prone to form a natural gathering point for activities. At the same time, the continuous gathering of commercial activities and people flows further attracts more businesses to move in and further strengthens the spatial aggregation of commercial facilities.

The degree of road integration in the new city station is higher than in the old one. The integration degree value is highest in the top four stations, with the number of commercial facilities equal to 3.2 times the integration degree value of the smallest four stations, indicating that accessibility has a specific spatial aggregation effect on commercial facilities. At the same time, the influence of accessibility on commercial facilities is significantly different in new and old urban areas. According to the current study, the old city metro station areas such as Erdaoqiao, Nanmen, and Beimen, which have low integration values, are, in fact, areas with a dense distribution of commercial facilities, which have not only well-developed infrastructures but also have traditional commercial centers with distinctive regional characteristics, such as the International Grand Bazaar, which is also a significant destination for tourists from all over the country. In terms of road network pattern, the road network of the old city consists of many winding and narrow side roads, resulting in generally low accessibility of the road network in the old city.

5. Discussion

Based on the historical commercial facilities POI data in Urumqi, this study explores the spatial distribution characteristics of commercial facilities around the stations of Urumqi Metro Line 1 and their influencing factors by using a comprehensive analytical method within the 800 m influence range of the metro stations.

In terms of changes in the density distribution of commercial facilities, after the construction of the subway, the kernel density value of commercial facilities within the scope of influence of all stations increased by 28–39%, indicating that the construction of the subway has played a promotional role in the further spatial aggregation of commercial facilities around the stations. The finding better supports the conclusion of the previous study, which found that subway construction favors the spatial concentration and diversification of various commercial facilities around stations [4,6,44,45]. However, the impact of metro construction on different types of businesses at stations in different urban areas in relevant studies is insufficiently understood, and the spatial development trend of different business facilities around stations under the influence of metro construction needs in-depth study. Our study goes further to analyze the impact of metro construction on different commercial facilities and spatial shaping, such as proposing that commercial facilities in old city stations grow more rapidly than those in new city stations under the influence of the metro and that some of the commercial facilities in stations under the influence of the metro present centripetal spatial agglomeration characteristics, while some of the commercial facilities in old city stations show the further weakening of the original centripetal aggregation pattern, such as catering, shopping, financial, living, and accommodation services, which indicates that the construction of the subway line has played a different spatial shaping role for the commercial facilities in the city with different basic conditions. The role of subway construction in shaping the space of commercial facilities varies under different underlying conditions in the city. This finding further complements recent research proposing that there are significant differences in the level of balanced commercial development and agglomeration between different types of station commercial space [46,47,48].

The change in the number of commercial facilities shows that after the construction of the subway, the number of commercial facility POIs within the influence of each subway station increased significantly, of which the number of accommodation, catering, and shopping facilities increased most significantly. This indicates that the construction of the subway has not only promoted the enrichment of each commercial sector but also has had a differentiated stimulating effect on different commercial facilities. The impact of subway construction on consumption-driven industries has been even more pronounced. These findings correspond to the conclusion that subway construction promotes commercial industries’ enrichment and vitality [5,49].

The results of influencing factors show that the station of Urumqi Metro Line 1 has a good spatial overlap with the high-density population area in the city, and the population density within the 800 m influence area of the metro station (5600 people/square kilometer) is higher than the average population density of the study area (3000 people/square kilometer), which better verifies the conclusion that the population factor is the main factor influencing the spatial distribution of the commercial facilities around the station [50,51]. The strong correlation between the change in the number of commercial facilities and some of the living infrastructures indicates that a sound infrastructure has a specific role in promoting the enrichment of commercial businesses around the site, but not all living infrastructures play a positive role in promoting the development of commercial businesses. This conclusion further validates the relevant findings of Lu Xiaoting et al. [52]. The accessibility factor has a significant spatial aggregation and commercial vitality enhancement effect on commercial facilities in Urumqi’s new city sites and a non-significant effect on commercial facilities in the old city sites. This finding not only validates the correlation between accessibility and the distribution of commercial space around the site, as concluded in previous studies [50,53,54], but also further illustrates the differential impact of accessibility on commercial space under different urban area conditions. Furthermore, the spatial syntax-based accessibility analysis method shows better adaptability in the new city area, and the analysis results in the old city area are less relevant to reality, which requires that the accessibility analysis integrate multiple methods to improve its accuracy [55].

This study has several limitations. First, the spatial distribution of commercial facilities around metro stations is influenced by multifaceted factors (e.g., political, economic, cultural, topographic, and land-use dynamics) [56]. However, only partial variables were analyzed, which may have introduced analytical bias. Second, the study selected limited POI data of six types of commercial facilities in Urumqi City in 2017 and 2022 and analyzed them comparatively to draw the above conclusions. Due to the discontinuousness of the annual data of the POIs of the commercial facilities used and the limited number of years spanning the period, the conclusions of the study may miss abnormal changes in the spatial characteristics of the commercial facilities in a particular year; at the same time, the limitations of the types of the commercial facilities selected may miss the special impact relationship between certain commercial facilities and metro construction, thus affecting the proposed targeted commercial facilities spatial layout strategy at the level of the totality of commercial facilities. Further research needs to be supported by years of continuous and complete commercial facility POI data to improve the reliability and credibility of the research conclusions. Third, the demographic impact analysis’s population data were limited to static data for 2022; longitudinal dynamic population data would enhance the robustness of the conclusions. Future studies will need more precise dynamic data to demonstrate the spatial distribution of population and commercial facilities around the subway, such as the hourly ridership of each subway station at different times of the day and the direction of ridership flow. Fourth, Metro Line 1 in Urumqi is relatively well-developed; the ongoing construction of Lines 2–4 may further reshape commercial spatial patterns. Future research should incorporate multi-line comparisons and longitudinal datasets to enhance the generalizability and completeness of findings.

6. Conclusions

This study investigates the distribution characteristics of commercial facilities around the stations of Urumqi Metro Line 1 through multi-method GIS analysis (kernel density estimation, standard deviation ellipse, and buffer zone analysis) of the historical POI data, supplemented by spatial syntax and correlation analysis to explore the relationship between the distribution of commercial facilities and the factors affecting them. The main conclusions are summarized as follows.

The construction of the subway has promoted the further clustering of businesses in the surrounding area, and this clustering effect varies depending on the conditions of the urban location where the station is located. Regarding the distribution of commercial facilities density, Urumqi Metro Line 1 high kernel density stations are mainly located in Erdaoqiao Station, Nanmen Station, Beimen Station, Nanhu Square, Nanhu Beilu, Daxigou, Sports Center, and other stations. As the above stations are located in the old city area in the urban core, with well-developed commercial facilities and infrastructural conditions, the value of the kernel density increased significantly after the construction of the subway.

There are spatial differences in the agglomeration characteristics of different commercial types. The overall commercial distribution after the subway construction shows the following two changes. First, after the construction of the subway, the original centripetal agglomeration of some commercial facilities has been further strengthened, such as sports and leisure facilities; Second, the characteristics of the commercial agglomeration have changed from a centripetal agglomeration before the construction of the subway to a piecemeal commercial agglomeration after the completion of the subway.

Subway construction will have different stimulating effects on commercial facilities in different infrastructural conditions, and the impact on consumption-driven industries will be more obvious. In terms of the number of POIs, before and after the completion of the subway, the number of POIs of all business types within the area of influence of the subway increased, and the number of POIs of accommodation, catering, and shopping facilities increased the most, as the rapid construction of the subway facilitated the development of all kinds of businesses, with the most significant impact on accommodation, catering, and shopping businesses. At the subway station level, Xinxing Street, North Gate, and South Gate stations have seen the most significant growth in the number of POIs, and the complete infrastructure conditions and unique regional landscape characteristics of these stations effectively promote the further aggregation of commercial businesses. Overall, the subway construction has effectively promoted the further development of various commercial sectors.

Since the subway construction, the direction of commercial development has been more consistent with the direction of the subway line. The plot of the standard deviation ellipses shows that the major axis of the SDEs represents the primary developmental orientation of geographic elements, with a mean angular deviation of 10.3° (p < 0.05). Meanwhile, 78% of new commercial POIs occur within the 800 m station buffer zone along the rail corridor.

Factors such as subway accessibility, population density, and living infrastructure all impact the distribution of businesses around the subway. Subway accessibility varies according to the subway station’s location and the city’s basic conditions, which have different impacts on commercial business. An increase in population density can increase commercial activity and promote the distribution of commercial clusters around the stations. Changes in the number of businesses have a significant positive correlation with changes in the number of public facilities and scenic spots, and the improvement of the above living infrastructure is conducive to the development of businesses around the subway.

This study provides a quantitative basis for the synergistic development of rail transit corridors and commercial spaces, guides the axial coupling of the land layout around the stations with the subway network, and especially emphasizes the priority of implanting consumption-driven businesses in areas with weak infrastructure to activate commercial potential. The study reveals the spatial coupling law of population density, public facilities, and commercial vitality, which provides a scientific path for city managers to formulate sustainable commercial density gradient control strategies based on the differences in station energy levels, as well as to enhance the sustainability of metro business through the optimization of living service facilities.