Evaluation of Sustainable Development Potential of High-Speed Railway Station Areas Based on “Node-Place-Industry” Model

Abstract

The development of the HSR station area is the result of the combined effect of the three elements of transport, place, and industry. This study introduces the industrial dimension and constructs the node-place-industry model to empirically analyze the development potential of station areas along the Hunan section of the Beijing–Guangzhou and the Shanghai–Kunming high-speed railway lines. The results show that (1) the development of the three spatial elements of the station area is mostly out of sync, and the node value has the highest fit with the integrated potential value of the station area; (2) there is a significant correlation between the magnitude of the combined potential of the station area and the site location, station class and time of development; (3) according to the results of the cluster analysis, it was found that most of the stations were in a state of disequilibrium, and the main reason was that the functional value of the place did not match with the value of industrial aggregation. In particular, the introduction of the industry dimension extends the NP model and establishes a tessellated analytical framework for station type classification, providing an interesting assessment tool for the sustainable development of transport hub areas.

1. Introduction

With the development of economic globalization and the increasing cross-border movement of global factors of production, such as capital, commodities, technology, and labor, between regions and even between continents, these demands have further stimulated the construction and upgrading of highways, high-speed railways, and large-scale national transport infrastructures, such as airports and ports. High-speed rail stations, airports, and intercontinental ports are increasingly becoming logistics centers and, through their interaction with the regions and cities in which they are located, promote the clustering of value-added industries, such as finance, commerce, and information, in transport hub areas. This promotes the integrated development of hub areas and urban hinterlands and leads to the reconstruction of urban spatial structure and center systems [1,2,3]. The deployment of high-speed rail services has been the most important innovation in global intercity travel in recent decades, with wide-ranging impacts on economic growth, transport accessibility, and social equity [4], with the development of high-speed rail in China very much represented. Since the opening of China’s first high-speed railway (HSR), the Qin-Shen Passenger Dedicated Line, in October 2003, China’s HSR has achieved leapfrog development. By the end of 2022, China’s HSR had a mileage of 42,000 km, accounting for 71% of the world’s total HSR mileage, with 1188 new HSR passenger stations. It is expected that by 2035, China’s HSR mileage will reach 70,000 km and the total number of new HSR passenger stations will exceed 2000. With its powerful spatial and temporal compression effect, HSR has reconfigured the regional spatial pattern and become an important lever to drive national and regional economic and social development [5]. The opening of high-speed rail can significantly improve accessibility to 32% of the areas along the route [4], and contribute to 8% of China’s GDP per capita [6]. HSR station areas are new economic forms with agglomeration and diffusion effects, consisting of core elements such as population, industry, and land [7]. The impact of HSR on cities and regions is realized through the rational organization of urban functions and spaces in and around HSR stations [8,9], which are becoming new key nodes in the country’s spatial development transformation.

As the timing of HSR construction in China is highly compatible with the accelerated development stage of urbanization, the governments of cities with stations have high expectations of urban spatial expansion and structural optimization through the development of HSR station areas. New cities and new districts have been planned and built based on the stations, making the HSR station area one of the hotspots for urban development studies at present [10,11]. However, in general, the planning and construction of HSR station areas in China are still in their initial stage, and the research on the spatial system of station areas and the integrated development of station cities is not yet in-depth. Most of the studies on the positioning of the functions and industrial development of the HSR station areas and the forecast of land use are based on qualitative analyses, empirical judgments, and foreign case studies. The lack of functional excavation of station area features and industrial system sorting makes the planning and design of station areas less scientific and less logical, resulting in the general lack of economic vitality and popularity of these station areas [12,13]. Most of the HSR station areas have been hollowed out due to the lack of industrial support during the planning and implementation process, and the current development status is lagging behind the planning expectation. There are even some uninhabited “ghost towns” due to the lack of relevant industry drive and industrial population import [14,15]. In addition, some of the HSR station areas tend to have a weak driving effect in the actual development process, especially in HSR station areas far away from cities, especially those located in villages and market towns, resulting in small added value [1]. According to overseas experience in the development and construction of high-speed railways and station areas, not all HSR station areas will receive major development. To this end, this paper provides a scientific, comprehensive, and objective evaluation of the development potential of HSR station areas, to reasonably determine the types of station areas, clarify the characteristics of station areas, and provide decision-making support for the formulation or optimal adjustment of station area planning and development programs, to promote the sustainable development of different types of station areas.

The node-place model (NP model) is a widely recognized method for evaluating and classifying the potential of transport hubs and has gone through three stages in different countries [16,17], from theoretical research to practical application and then to model expansion, and its research areas have also shown a diversified trend. In recent years, most studies on transport interchange potential assessment have focused on rail transit stations under the TOD development model [18,19,20,21], with little attention paid to HSR stations. Some scholars have found that HSR can promote the agglomeration of service industries by reducing transport costs and that the station areas of the world’s developed HSR are dominated by tertiary industries [22]. HSR stations are geographic spaces with a distinct circle structure, combining economic, industrial, and spatial concepts. The combination of HSR stations and industrial space has become an emerging field of interest for many disciplines, including economic geography, urban planning, and transport economics [7].

The NP model is a widely used and effective evaluation method, but its evaluation is based on the balanced development of traffic node value and place function value of transport interchange but lacks the analysis and quantification of industrial aggregation value. At the same time, considering that the industrial aggregation analysis is the basis for determining the functional positioning of the station area, it has a great impact on the development potential of the station area. Therefore, to achieve the goal of sustainable development of the HSR station area, it is particularly crucial to introduce the industrial dimension into the station area potential evaluation model.

Based on the perspective of station–city interaction, this paper introduces the industry dimension to the NP model for the first time, adds evaluation indicators such as the development level of station cities and the potential for industrial clustering in station areas, and builds the “node-place-industry” model (NPI model) and a new evaluation index system. At the same time, the NPI model was applied to empirically rate the development potential of 18 station areas in the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines and to determine the type of the station, the original NP model has been innovatively extended at both the theoretical and practical levels. This study aims to answer the following research questions. (1) What is the current level of sustainable development potential of the 18 station areas in the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines from the perspective of the coupling of node, place, and industrial values in HSR station areas? (2) How does the governmental management of the station locations cope with the imminent risks of sustainable development in HSR station areas?

This paper is divided into six sections. Based on this section, Section 2 discusses the theoretical background of the Node-place model and its extended application and explores the existing methods and experiences in evaluating the development potential of HSR station areas. Section 3 introduces the industry dimension, constructs the node-place-industry model and evaluation index system, and uses hierarchical analysis, principal component analysis, and hierarchical cluster analysis to assess the development potential of the 18 station areas in the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railways. Section 4 reveals the results of the node-place-industry model in the empirical study. Section 5 summarizes the research process and the results of the data analysis and presents the innovation points of this study. Section 6 concludes and summarizes the whole text.

2. Literature Review

2.1. The Node-Place Model and Its Extended Applications

Bertolini proposed the Node-place model (NP model), which suggests that station areas contain both node (transport) and place (function) values [8], and that these two values are mutually compatible and supportive of the sustainable development of station areas. The y-axis of the NP model represents node values, reflecting transport attributes. The x-axis represents place values, reflecting functional attributes. Based on this model, transport interchange areas can be classified into five types: independent, balanced, stressed, unbalanced node, and unbalanced place (Figure 1). When the accessibility or land use functional status changes, the development status in which the station area is located changes accordingly. The core idea is to achieve sustainable development of the area by increasing the degree of coupling of node value and place value in the hub area. The NP model is widely used in the assessment of transport hub areas. It can effectively identify the development potential of transport hub areas and classify stations [23]. On this basis, through assessment and planning guidance, the state of the hub area will be induced to move in the direction of sustainable development close to the diagonal, which is the significance of conducting an assessment of the development potential of the HSR station area, providing a theoretical framework for the development of transport hub areas.

In recent years, foreign scholars have continued to add and explore more dimensions to the NP model, and these useful attempts have expanded the ideas for station area assessment from specific perspectives. For example, Groenendijk et al. added node quality to the NP model from a traveler’s perspective, leading to the development of the node-place-experience (NPE) model [24]. Zhang et al. constructed the node-place-design (NPD) model, dividing London’s rail stations into five categories [25]. Pezeshknejad et al. used design indicators as a third dimension in their study of BRT systems [26], while Dou et al. derived from network theory to classify metro stations and construct the node-place-network (NPN) model [27]. Su constructed the node-place-function (NPF) model [28], which projects the indicators of the node, place, and function dimensions into a three-dimensional coordinate system and classifies each dimension into high, medium, and low categories, generating 27 cubes in the three axes, each of which represents a specific type of TOD. Yang et al. expanded the node-place model into a four-dimensional node-place-design-vibrancy (NPNV) model in their study of rail transit [29], and used Guangzhou and Shenzhen as the subjects of their comparative study, ultimately concluding that Shenzhen’s TOD performed better than Guangzhou’s in general. Zhou et al. developed a node-place-association (NPA) model to empirically analyze Hong Kong’s Metro stations, taking into account social interactions [16].

The NP model is a powerful analytical tool for describing station areas from the interaction between transport and land use development, but it also has obvious flaws. That is, the starting point of its evaluation is based on whether the development of station areas is in a balanced state or not, and the five station area types it identifies are difficult to effectively guide the future development of a large number of transport hubs areas with multiple levels. The useful approach of the NP extension model expands the ideas for the assessment of the HSR station area from a specific perspective, but the extension of the model neglects the study of spatial mechanism and lacks the consideration of the industrial agglomeration level, so this paper introduces the industrial dimension to extend the NP model.

2.2. Evaluation of the Development Potential of the HSR Station Area

Based on the NP model, a number of scholars have quantitatively assessed the sustainability of the HSR station area. Reusser et al. applied the NP model, taking into account both station connectivity to other places and activity around the station, node value was measured using indicators such as the direction of train service, frequency of train service, number of stations within 20 min of travel, number of directions of other public transport, daily frequency of other public transport, and distance to the nearest motorway entrance, number of cycle lanes, etc., and place value was measured using number of people, number of workers per economic vector, and degree of functional mix. Extending the NP model through expert questionnaires and reserve grid interviews, and finally classifying all railway station points in Switzerland through cluster analysis, a non-linear relationship between node value and place value was observed and can be used for the pre-screening of stations to be optimized [30]. Ortuno-Padilla et al. applied the NP model to evaluate the development potential of 21 railway stations around Mianyang City, Sichuan Province, a less-developed region of China, providing effective data support for the application of the model through a detailed field survey [31]. They selected railway accessibility, urban transport accessibility, intercity bus accessibility, car accessibility, bicycle and electric motorbike accessibility, and distance from the city center to measure node value, and labor force distribution and degree of population mixing as place value to classify the sites into five categories. Babb et al. used the NP model to assess the transport function and land use status of Perth railway station [32]. A three-dimensional analysis of node, place, and background traffic for stations within and beyond the motorway median and stations on conventional lines, taking 43 metrics to measure them, found that achieving a balance of node and place values associated with motorway median stations was not realistic and that a better policy would be to emphasize nodal functions. Nie Jing used the NP model to develop an evaluation model of passenger flow to evaluate the current development status of each HSR station area [33]. Using regression cluster analysis to calculate the correlation between the average annual development of station areas and urban development, urban dynamics, station class, transport accessibility, station area land, and other factors, to derive the dominant factors that actually affect the development of HSR station areas in large, medium, and small cities, respectively. Kim et al. quantified node value in terms of station accessibility and availability of train connections, and place value in terms of population size, number of employees, and multifunctionality of space use in the station area [34]. The assessment analyzed the performance of stations and network performance, observed the rail network in terms of the degree of centrality and meso-centrality, and ultimately proposed three guidelines for optimizing the train station network: integrating resilience practices, coordinating a high level of cooperation between stations, and creating a culture of network risk management. Gui et al. used the NP model to calculate the node and place values of 24 HSR stations along the Beijing-Shanghai high-speed railway line [35]. It was found that only very few stations were in equilibrium and through correlation analysis it was learned that node value was related to city size and place value was not strongly linked to city size. Wei et al. assessed the development potential of HSR station areas at three spatial scales: regional, urban, and station areas. The development potential of 123 HSR station areas in the Yangtze River Delta region was evaluated by characterizing the node value in terms of accessibility, service level, and network centrality of HSR station areas, and the place value in terms of vitality value, facility density and land use diversity [36].

In addition to the NP model, studies on the evaluation of the development potential of HSR stations contain other qualitative and quantitative types [37]. And quantitative evaluation is the main focus, with the commonality being a general focus on the economic development potential of stations. In terms of qualitative evaluation, for example, Xiong Wei et al. analyzed the industrial planning and layout of Taizhou City, a key city along the Hang Shaotai Railway line, and then explored the potential for comprehensive land development in and around Taizhou Station [38]. In terms of quantitative evaluation, for example, Liu Hua identified two aspects, namely the economic benefits of the HSR line and the planning design of the HSR station, as the two main dimensions affecting the development potential of HSR station areas. The DEA method and fuzzy evaluation method were used to quantify the two types of indicators, respectively, and the comprehensive evaluation results of the development potential of 17 stations along the Beijing-Shanghai high-speed railway line were obtained [39]. Di Matteo et al. used the Milan-Bologna HSR Corridor in northern Italy as a study to assess the economic performance of HSR in the service and manufacturing sectors through the propensity score matched difference method (PSM-DID) [40]. Zou et al. constructed a development potential evaluation index system from three aspects: the site rationality of the station, the service capacity of the station, and the development level of the cities where the stations are located. The development potential ofnine9 representative HSR station areas of different grades along the Wu-Guangdong Passenger Dedicated Line was evaluated [41].

With the popularity of big data and new technologies, some scholars have used nighttime lighting data to analyze the development potential of HSR station areas. For example, Du et al. analyzed the spatial and temporal distribution of brightness in 980 HSR stations in China based on nighttime lighting data and classified the stations into four categories based on their brightness and relative brightness during HSR operations [42]. Point of interest (POI) data are also widely used to assess the potential of the HSR station area [43], and post-use assessment data, such as public review data, is also used to assess the spatial vitality of the HSR station area [44]. Overall, open big data offers new ways to select and calculate indicators for assessing the potential of HSR station areas [36].

The above studies have not yet reached a consensus on the selection of evaluation indexes, mainly adopting the intensity and diversity of traffic supply to measure the value of node, but the division of categories is not clear. The intensity and diversity of urban activities in the station area are used to characterize the value of the place, but fewer qualitative indicators are included, and overall, the evaluation indicator system is not constructed from the perspective of integration of industry and city and integration of station and city. Moreover, most of the assessment results are to categorize the research subjects into a few types and make recommendations, making it difficult to break out of the typological framework of the NP model. Based on the spatial characteristics analysis of high-speed railway station areas, this paper will reconstruct a brand-new evaluation index system of station area development potential, set up three-level indexes, increase quantitative data, and ultimately systematically classify the types of station areas by combining the evaluation results.

3. Methodology

3.1. Potential Evaluation Model

3.1.1. Correspondence between Spatial Elements and Potential Assessment Dimensions of the HSR Station Area

Based on the high accessibility of the HSR station to gather a large number of factor flows, the HSR station has a direct or indirect economic impact on the surrounding areas, mainly attracting modern service industries and advanced manufacturing industries to move in. HSR station area has gradually become a new type of industrial agglomeration area in the city. The spatial system of the HSR station area consists of three main elements: transportation space, physical space, and industrial space [45], which correspond to the three dimensions of the assessment of the development potential of the station area. The transport space includes the HSR station collection and distribution mode and the interface between the HSR station and inner and outer city traffic, representing traffic node values, corresponding to node dimension. Physical space is the concrete land use and spatial expression of the HSR station area as an urban functional area, reflecting the functional value of the place and corresponding to the dimension of the place. The industrial space includes the type and the scale of modern services and advanced manufacturing, and the spatial layout of the HSR station area, reflects the value of industrial aggregation and corresponding to the industrial dimension

The three spatial elements of the HSR station area are interrelated and influence each other, with changes in any one of them causing a corresponding response in the other two. Among them, the transport node is the catalyst for the development of the station area, the physical place is the land and spatial carrier for the development of the station area, and the industrial aggregation is the feedback channel between the coordination of the transport supply and the function of the place and is also the result of the development of the station area. A rational spatial structure of the station area has a significant effect on the concentration of industry and people and will greatly enhance the function of the transport facilities in the HSR station area itself. Industrial upgrading and spatial optimization are all dependent on the support of the station area transport network [46]. Thus, it is only through the synergistic development and interactive feedback of the three elements that the greatest synergy can be generated to drive the development of the HSR station area (Figure 2), which is an important guarantee for the high-quality development and sustainable development of the station area.

3.1.2. NPI Model Construct

The NP model for evaluating the development potential of HSR station areas is more mature in European countries, and relevant studies abroad have mostly used model expansion as an entry point to improve the indicator system, but no scholars have yet incorporated the industry dimension into the NP model. Compared with foreign countries, the application of the NP model in China is relatively small. Most of the relevant studies in China follow the original NP model, and only a few scholars have extended the model. There are problems such as rough delineation of station area types, one-sided evaluation indicators, single application scenarios, and insufficient support for emerging data.

The three spatial elements of the HSR station area are indispensable for the development of the station area, and industrial support is particularly crucial. If the station area lacks relevant industries and industrial population introduction, it will be in the situation of “having a city without production” and lose the ability for sustainable development. Based on this, the traditional NP model is expanded to include the industrial dimension, and the node-place-industry (NPI) model is constructed (Figure 3). Three main dimensions are included in this model to evaluate the potential of the station area: (1) traffic indicators, reflecting the accessibility of the station area by different modes of transport and emphasizing the value of traffic; (2) land-related indicators, characterizing the degree of site size and mixed land use development around the station area, emphasizing the value of urban functions; (3) economic indicators, reflecting the state of aggregation of industrial development in the station area and emphasizing the value of industrial aggregation.

Drawing on the ‘Rubik’s Cube’ analytical framework proposed by Su et al. which projects the node-place-industry dimension into a three-dimensional coordinate system [28], the x, y, and z axes are set as node, place, and industry values, respectively. Taking any vertex as an example, the three adjacent faces consist of node-place, node-industry, and place-industry sub-models, respectively. At the same time, each axis can be classified into three levels of high (H), moderate (M), and low (L) according to the corresponding quantitative intensity, then the potential assessment based on the NPI model can be abstractly understood in the form of a Rubik’s Cube similar to a 3 × 3 × 3. Tertiary intensity quantification and station type determination was carried out as follows. Firstly, outliers were observed in the normalized data by Z-score, and manual judgment was made. After excluding the outliers, the individual indicators are divided into the remaining data using the quantile method. The standardized data within the study area are sorted in descending order, with data below the first quartile as low, data between the first and third quartiles as middle, and data above the third quartile as high, and the three metrics together combined to form the 27 station area types. A total of 27 small cubes were generated, encompassing all station types, with each cube representing a specific station potential type (Figure 4). According to the combined relationship between node, place, and industry value of station areas, HSR stations can be classified into 7 major categories and 27 sub-categories (

Table 1. HSR Station type classification: H, high degree; M, moderate degree; L, low degree.

Number	Broad Category	Sub-Category	Node	Place	Industry	Implications
1	Dependent	NPI Dependent	L	L	L	Low-level interdependence of NPI.
2	Class dependent	Node Lightly Dominated Class Dependent	M	L	L	Moderate development of N and low level of interdependence of PI.
3		Node Heavily Dominated Class Dependent	H	L	L	Developed N, low level of interdependence of PI.
4		Place Lightly Dominated Class Dependent	L	M	L	Moderate development of P and low level of interdependence of NI.
5		Place Heavily Dominated Class Dependent	L	H	L	Developed P, low level of interdependence of NI.
6		Industry Lightly Dominated Class Dependent	L	L	M	Moderate development of I and low level of interdependence of NP.
7		Industry Heavily Dominated Class Dependent	L	L	H	Developed I, low level of interdependence of NP.
8	Balanced	NPI Balanced	M	M	M	Mutually balanced NPI.
9	Class Balanced	Node Deficient Class Balanced	L	M	M	Lack of N development with PI balanced.
10		Node Developed Class Balanced	H	M	M	Well-developed N with a balanced mix of PI.
11		Place Deficient Class Balanced	M	L	M	Lack of P development with NI balanced.
12		Place Developed Class Balanced	M	H	M	Well-developed P with a balanced mix of NI.
13		Industry Deficient Class Balanced	M	M	L	Lack of I development with NP balanced.
14		Industry Developed Class Balanced	M	M	H	Well-developed I with a balanced mix of NP.
15	Stressed	NPI Stressed	H	H	H	NPI in mutual tension.
16	Class Stressed	Node Deficient Class Stressed	L	H	H	Lack of N development and mutual tension between PI.
17		Node Balanced Class Stressed	M	H	H	Moderate development of N and mutual tension between PI.
18		Place Deficient Class Stressed	H	L	H	Lack of P development and mutual tension between NI.
19		Place Balanced Class Stressed	H	M	H	Moderate development of P and mutual tension between NI.
20		Industry Deficient Class Stressed	H	H	L	Lack of I development and mutual tension between NP.
21		Industry Balanced Class Stressed	H	H	M	Moderate development of I and mutual tension between NP.
22	Unbalanced	Node Heavy and Place Light Dominant	H	M	L	Developed N, moderate P, lack of I development.
23		Node Heavy and Industry Light Dominant	H	L	M	Developed N, moderate I, lack of P development.
24		Place Heavy and Node Light Dominant	M	H	L	Developed P, moderate N, lack of I development.
25		Place Heavy and Industry Light Dominant	L	H	M	Developed P, moderate I, lack of N development.
26		Industry Heavy and Node Light Dominant	M	L	H	Developed I, moderate N, lack of P development.
27		Industry Heavy and Place Light Dominant	L	M	H	Developed I, moderate P, lack of N development.

), providing a basis for the subsequent grading and classification of the development potential of the 18 HSR stations along the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines.

3.2. Selection and Measurement of Model Evaluation Indicators

The NPI model focuses on evaluating the value of node, place, and industry in the station area, with the following system of evaluation indicators. The connotation of each evaluation indicator, the calculation of the evaluation value, and the source of the underlying data are shown in

Table 2. NPI model evaluation indicators and calculation methods.

Category	Dimension	Indicator	Calculation Method	Data Sources
Node	External traffic service capacity	Station design scale	A number of 2 stations and 4 lines (County Stations) 30; 3 stations and 6 lines, 3 stations and 7 lines, 4 stations and 8 lines (Prefecture Stations) 50–60; 5 stations and 11 lines, 8 stations and 19 lines (Prefecture Hub Stations) 70–80; 13 stations and 28 lines (Provincial Capital Hub Stations) 100.	Baidu encyclopedia accessed on 12 August 2022 https://baike.baidu.com/
		Station passengers’ flow	Average daily passenger traffic during the Spring Festival in 2023, average daily 0–2000 (40–59), average daily 2001–7000 (60–69), average daily 7001–10,000 (70–79), average daily 10,001–49,999 (80–89), average daily 50,000 and above (90–100).	Official press releases
		Train frequency	Total number of daily departures or stops of high-speed trains, city trains, and trains.	Railway 12306 official website accessed on 12 August 2022 https://www.12306.cn/index/
		Distance of the station from the nearest motorway interchange	Route distance to the nearest motorway interchange.	Gaode Map’s official website accessed on 12 August 2022 https://ditu.amap.com/
	Capacity to connect with city traffic	Number of public transport lines	Number of public transport lines (including rail lines) within a 500 m radius of the HSR station.	Gaode Map’s official website accessed on 20 December 2022 https://ditu.amap.com/
		Number of public transport stations	Number of public transport stations (including rail stations) within a 500 m radius of the HSR station.
		Public transport time to the city	Fastest time by public transport from the HSR station to the city center.
		Driving time to downtown	Fastest drive time from the HSR station to the city center.
Place	The station’s development location	Spatial locational characteristics	Relative distance and location relationship between the HSR station and the center of the current urban built-up area: Rural (25–40), Bazaar (40–55), Suburban (55–70), Peripheral (70–85), Urban (85–100).	Gaode Map’s official website accessed on 20 December 2022 https://ditu.amap.com/
		Harmonization of the station with urban space	Harmonization of the HSR station with the main directions of urban spatial expansion: Compliance (100–80), Deviation (79–60), Departing (59–40)
		Spatial compactness of the urban form in which the station is located	$\frac{2\sqrt{\pi A}}{P}$, where A is the built-up area and P is the city profile perimeter.
	Functional orientation	Functional orientation of the station area	Gateway to urban subcentres 100–90; Urban Fringe HSR Custers 80–70; Industrial Park Integrated Service Areas 70–60; Suburban HSR Integrated Clusters 60–50; Stand-alone HSR Character Towns 50–40.	Regulatory detailed planning or urban design scheme for the area where the station is located
	Diversity of land use in the station area	Proportion of land for residential	The proportion of planned residential land area to the area of the HSR station area.	Tiantu Official Website accessed on 20 December 2022 https://www.tianditu.gov.cn/
		Proportion of commercial service facilities	The proportion of planned commercial service facilities’ land area to the area of the HSR station
		Proportion of land for public service facilities	The proportion of planned public service facilities’ land area to the area of the HSR station.
		Proportion of land for integrated transport facilities	The proportion of planned integrated transport facilities land area to the area of the HSR station.
		Proportion of land for industrial and logistics storage	The proportion of planned industrial and logistics storage land area to the area of the HSR station.
Industry	Development level of station cities	GDP of the city or county	GDP of the city or county where the HSR station is located.	Statistical Bulletin on National Economic and Social Development
		Population size of the city or county	The overall population of the city or county where the HSR station is located.	Seventh National Census Data of China
		Per capita GDP of the city or county	Gross domestic product per capita of the city or county where the HSR station is located.	Statistical Bulletin on National Economic and Social Development
		Public service level of the city or county	Local government public finance revenues of the city or county where the HSR station is located.
		Industrial structure of the city or county	The ratio between the primary sector (A), secondary sector (I), and tertiary sector (S) of the city or county where the HSR station is located, Pre-Industrial (A > I) 40–59, Early Industrial (A > 20%, A < I) 60–69, Mid-Industrial, (A < 20%, I > S) 70–79, Late Industrial (A < 10%, I > S) 80–89, Post-Industrial (A < 10%, I < S) 90–100.
	Potential for industrial agglomeration	Secondary industry enterprises’ density	The ratio of the number of secondary industry enterprises in the HSR station area to the planned land area of the HSR station area.	Gaode Developer Platform accessed on 20 December 2022 https://lbs.amap.com/
		Productive services enterprises’ density	The ratio of the number of productive service enterprises in the HSR station area to the planned land area of the HSR station area.
		Living services enterprises’ density	The ratio of the number of living services enterprises in the HSR station area to the planned land area of the HSR Station area.

.

• Node value reflects the ability of the HSR station to serve external traffic and connect with urban traffic. The external service capacity of the station is characterized by the design size of the station, the volume of passengers of the station, the frequency of trains, and the distance to the nearest high-speed entrance/exit to the station. The number of public transport routes, the number of public transport stations, and the time taken to reach the city center are used to characterize the level of connectivity of the station to the city.
• Place value reflects the station’s development location, the functional positioning of the HSR station area, and the degree of mixing of land use, emphasizing the integration of the station city. Macro measures the spatial coherence between the station and the city, meso measures the level of decision-making on functional positioning, and micro measures the degree of composite land use in the station area. Relevant indicators may include the station’s development location (classified into five categories: Rural, Bazaar, Suburban, Peripheral, and Separation according to the path distance from the center of the built-up area, three categories: compliant, deviated and urban according to the coordination with the direction of urban spatial expansion, and the compactness of the urban spatial form in which the station is located), the functional target positioning of the station area (gateway to urban sub-centers, high-speed rail zones, high-speed rail clusters, stand-alone high-speed rail character towns, and peripheral urban transport hubs), and the degree of functional complexity of the various types of the planned sites in HSR station areas.
• Industrial value reflects the development level of the city where the station is located and the potential for industrial agglomeration and development in the station area, with emphasis on industry–city integration. Industrial value reflects the level of development of the city where the station is located, and the potential for industrial agglomeration in the station area. At a macro level, it reflects the development level of the economy, population, and industrial structure of the city (county) in which the station area is located. At a micro level, it measures the density of enterprises in the secondary industry, production services, and domestic services within the station area.

3.3. Research Subjects

Hunan Province is in the transition zone between the eastern coastal region and the western region, with the pivotal position of bearing the east and connecting the north and the south. The Beijing–Guangzhou high-speed railway line and the Shanghai–Kunming high-speed railway line were opened early, with important lines and a high number of stations set up in Hunan Province (Figure 5). This paper takes 18 stations along the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines as the object of study. The planning area of the HSR new district corresponding to each station is taken as the study area. A summary of the stations is shown in

Table 3. Overview of stations along the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed rail lines *.

Number	HSR Station	Location	Design Scale	Functional Positioning of the HSR Station Area
1	Yueyangdong	Edge of Yueyang City	3 stations, 7 lines	Integrated Urban Sub-centre
2	Miluodong	Edge of Miluo Industrial Park, Yueyang City	2 stations, 4 lines	Industrial Park Integrated Service Area
3	Changshanan	Edge of Changsha City	13 stations, 28 lines	Commercial and Business-Oriented Sub-centre
4	Zhuzhouxi	Edge of Zhuzhou City	3 stations, 7 lines	Commercial and Business-Oriented Sub-centre
5	Hengshanxi	Rural area of Kaiyun Town, Hengshan County, Hengyang City	2 stations, 4 lines	HSR Tourist Town
6	Hengyangdong	Edge of Hengyang City	5 stations, 11 lines	Commercial and Business-Oriented Sub-centre
7	Leiyangxi	Edge of Leiyang District, Hengyang City	2 stations, 4 lines	Integrated Urban Sub-centre
8	Chenzhouxi	Edge of Chenzhou City	3 stations, 6 lines	Urban Edge HSR Cluster
9	Lilingdong	Far suburb of Liling District, Zhuzhou City	2 stations, 4 lines	Urban Edge HSR Cluster
10	Xiangtanbei	Edge of Jiuhua Economic Development Zone, Xiangtan City	3 stations, 7 lines	Industrial Park Integrated Service Area
11	Shaoshannan	Edge of Shaoshan District, Xiangtan City	2 stations, 5 lines	Tourism-serving Sub-centre
12	Loudinan	Periphery of Loudi City	4 stations, 8 lines	Commercial and Business-Oriented Sub-centre
13	Shaoyangbei	Rural area of Pingshang Town, Xinshao County, Shaoyang	2 stations, 4 lines	HSR Industry and Trade Town
14	Xinhuanan	Rural area of Yangxi Town, Xinhua County, Loudi City	2 stations, 4 lines	Publishing and Printing Town
15	Xupunan	Rural area of Beidouxi Town, Xupu County, Huaihua City	2 stations, 4 lines	HSR Tourism Town
16	Huaihuanan	Edge of Huaihua City	8 stations, 19 lines	Integrated Urban Sub-centre
17	Zhijiang	Periphery of Zhijiang County, Huaihua City	2 stations, 4 lines	Urban Fringe Trade and Logistics Cluster
18	Xinhuangxi	Far suburb of Xinhuang County, Huaihua City	2 stations, 4 lines	Suburban HSR Cluster

* The Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines were opened for operation in December 2009 and December 2014, respectively.

.

3.4. Data Processing

After obtaining the basic data, the evaluation index system is first constructed using hierarchical analysis to determine the weights of the criterion layer and sub-criterion layer. Secondly, the data for each indicator term were standardized by the extreme value method to eliminate the effect of different magnitudes. To determine the degree of influence of each indicator item on the development of the HSR station area, principal component analysis was used to pre-process the secondary indicators to produce a set of linearly unrelated variables, reduce the number of clustered variables to obtain a comprehensive evaluation index of the development potential of the station area, determine the ranking of the development potential of each station area, and then conduct a hierarchical cluster analysis. Finally, the 18 stations in the Hunan Section of the Beijing–Guangzhou and Shanghai–Kunming High-Speed Rail lines were classified into eight station types according to the classification results by applying them to the NPI model.

3.4.1. Analytic Hierarchy Process

First introduced in the 1970s by American operations researcher TL Saaty [47], hierarchical analysis is a decision analysis method that integrates qualitative and quantitative indicators. It is capable of decomposing complex problems into multiple constituent elements, building structural models with multiple levels of progression, and constructing judgment matrices for two-way comparisons. Weights are determined by solving the eigenvectors of the relative importance judgment matrix and consistency tests are performed on the hierarchical ranking.

The first step is to construct the judgment matrix, while the second step is to solve for the weight magnitude using the square root method. Let $\eta$ be the eigenroot of A and $W={(W_1,W_2,\cdots ,W_n)}^T$ be the eigenvector corresponding to the eigenroot. First, calculate the nth root ${\stackrel{-}{W}}_i$ of the product of the elements of each row of the judgment matrix A (Equation (1)), then the normalization (Equation (2)) is followed by a consistency test and finally, a hierarchical total ranking is generated.

$${\stackrel{-}{W}}_i=\sqrt[n]{{\prod }_{j=1}^n{a}_{ij}}i=\mathrm{1,2},\cdots ,n$$

(1)

$$W_i=\frac{{\stackrel{-}{W}}_i}{{\sum }_{i=1}^n{\stackrel{-}{W}}_i}i=\mathrm{1,2},\cdots ,n$$

(2)

3.4.2. Normalization

The normalization of data is the scaling of data so that they fall into a specific interval and the effect of the magnitude is eliminated. We standardized the data for the 25 indicator items in the HSR station area using the extreme value method (Equation (3)), scaling all indicators to a range of 0 to 1.

$$X_{ij}^{\prime }=\begin{array}{cc}\frac{X_i-Min{(X}_{ij})}{Max\left(X_{ij}\right)-Min{(X}_{ij})}& \\ & \end{array}$$

(3)

In the formula, $X_{ij}^{\prime }$ is the value of indicator i for the HSR station area of j. For all station areas, $Max\left(X_{ij}\right)$ is the maximum value of the indicator value of i for the HSR station area of j, and $Min{(X}_{ij})$ is the minimum value of the indicator value of i for the HSR station area of j.

3.4.3. Principal Component Analysis

Principal component analysis was introduced by Karl Pearson for non-random variables [48], and Hotelling extended the method to the random vector case [49]. The dimensionality of the data space under study is reduced by transforming a set of potentially correlated variables into a set of linearly uncorrelated variables through an orthogonal transformation. The final evaluation value is obtained by weighting and summing the principal components, with the weight being the variance contribution of each principal component.

3.4.4. Hierarchical Clustering Analysis

Hierarchical clustering analysis is an algorithm that analyzes data at different levels based on the similarity between clusters, resulting in a tree-like clustering structure [50,51]. In contrast to k-means clustering analysis, hierarchical clustering analysis does not require the k-values to be determined in advance; the basic idea is to calculate the similarity matrix of the data set. Assuming that each sample point is a cluster, the two clusters with the highest similarity are merged and the similarity matrix is updated, and the merging process is repeated until the number of clusters is one.

Calculation of inter-cluster similarity using average distance. First, assume that the inter-sample distance is dist($P_i$, $P_j$), where $P_i$, $P_j$ are any two samples, $\left|C1\right|$, $\left|C2\right|$ denotes the number of samples in the cluster class, respectively, the distances for cluster classes $C1$ and $C2$ are calculated as shown in Equation (4).

$$\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}\left(\left|C1\right|, \left|C2\right|\right)=\frac{1}{\left|C1\right|·\left|C2\right|}\sum _{P_i\in C1,P_j\in C2}\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}(P_i,P_j)$$

(4)

Squared Euclidean distance is then used to calculate the distance between samples and let the samples be n-dimensional. The distance is calculated as shown in Equation (5).

$$\mathit{dist}(P_i,P_j)={\displaystyle \sum _{k=1}^n}{(P_{ik}-P_{jk})}^2$$

(5)

4. Results

4.1. Results of the Analytic Hierarchy Process

Based on the concept of the interactive development of transport space, physical space, and industrial space in the HSR station areas, a three-level framework system comprising a goal level, a criteria level, and an indicator level has been constructed. The goal level is the evaluation of the development potential of the HSR station area, and the criteria level contains three core subsystems: node value, place value, and industry value. The node value consists of two sub-criteria levels, namely the ability to serve external traffic and the ability to connect to urban traffic. The place value consists of three sub-criteria levels: the station’s development location, the function and positioning of the station area, and the diversity of land use in the station area. The industrial value consists of two sub-criteria levels, namely the level of development of the city with a HSR station and the potential for industrial clustering in the HSR station area. The indicator levels are the 25 indicator items corresponding to the criteria levels at each level. The final node, place, and industry value weights were obtained as 0.493, 0.196, and 0.311, respectively (Figure 6).

4.2. Results of the Principal Component Analysis

The normalized data were entered into the IBM SPSS Statistics 26 platform for factor analysis under the dimensionality reduction module. According to the principle of extracting principal components with characteristic roots greater than 1, two node value principal components, three place value principal components, and three industry value principal components can be extracted, respectively. A table of component matrices and the total variance explained was obtained, and then the loading coefficients were calculated by dividing the elements of the component matrix by the standard deviation of the original variables (

Table 4. Component matrix of node, place, and industry values.

Node			Place				Industry
Indicator	Ingredient		Indicator	Ingredient			Indicator	Ingredient
	1	2		1	2	3		1	2	3
C1	0.928	−0.152	C9	0.749	0.594	0.015	C18	0.726	−0.518	−0.296
C2	0.891	−0.153	C10	0.823	0.473	0.102	C19	0.569	−0.089	−0.791
C3	0.887	−0.274	C11	0.055	−0.416	0.590	C20	0.607	−0.664	0.370
C4	0.839	0.169	C12	0.865	0.374	0.055	C21	0.805	−0.176	0.237
C5	0.773	−0.469	C13	0.170	−0.160	−0.843	C22	0.443	−0.441	0.263
C6	0.519	0.024	C14	−0.786	0.407	−0.065	C23	0.553	0.711	−0.053
C7	0.352	0.838	C15	0.213	−0.437	0.540	C24	0.450	0.802	0.184
C8	0.554	0.780	C16	−0.533	0.686	0.226	C25	0.687	0.601	0.117
			C17	−0.428	0.768	0.182

,

Table 5. Total variance explained for node, place, and industry values.

Ingredient	Initial Eigenvalue			Sum of Squares of Extracted Loads
	Aggregate	Percentage of Variance	Cumulative Percent	Aggregate	Percentage of Variance	Cumulative Percent
1	4.444	55.547	55.547	4.444	55.547	55.547
2	1.681	21.013	76.560	1.681	21.013	76.560
3	0.824	10.304	86.864	-	-	-
4	0.429	5.368	92.232	-	-	-
5	0.365	4.568	96.800	-	-	-
6	0.141	1.763	98.563	-	-	-
7	0.078	0.980	99.543	-	-	-
8	0.037	0.457	100.000	-	-	-
9	3.148	34.982	34.982	3.148	34.982	34.982
10	2.332	25.910	60.892	2.332	25.910	60.892
11	1.453	16.148	77.040	1.453	16.148	77.040
12	0.787	8.744	85.784	-	-	-
13	0.507	5.637	91.420	-	-	-
14	0.303	3.365	94.785	-	-	-
15	0.241	2.673	97.458	-	-	-
16	0.128	1.423	98.881	-	-	-
17	0.101	1.119	100.000	-	-	-
18	3.044	38.046	38.046	3.044	38.046	38.046
19	2.452	30.653	68.699	2.452	30.653	68.699
20	1.026	12.826	81.525	1.026	12.826	81.525
21	0.801	10.016	91.542	-	-	-
22	0.298	3.724	95.266	-	-	-
23	0.204	2.545	97.810	-	-	-
24	0.154	1.927	99.737	-	-	-
25	0.021	0.263	100.000	-	-	-

and

Table 6. Indicator loading coefficients and principal factor contributions of principal component analysis.

Node			Place				Industry
Indicator	First Principal Component	Second Principal Component	Indicator	First Principal Component	Second Principal Component	Third Principal Component	Indicator	First Principal Component	Second Principal Component	Third Principal Component
C1	0.440	−0.117	C9	0.422	0.389	0.013	C18	0.531	−0.036	−0.262
C2	0.367	−0.362	C10	0.464	0.310	0.085	C19	0.327	0.035	−0.723
C3	0.421	−0.211	C11	0.031	−0.272	0.490	C20	0.497	−0.116	0.399
C4	0.246	0.018	C12	0.488	0.245	0.045	C21	0.437	−0.151	0.057
C5	0.423	−0.118	C13	0.096	−0.105	−0.700	C22	0.346	−0.003	0.476
C6	0.398	0.131	C14	−0.443	0.267	−0.054	C23	0.101	0.570	0.014
C7	0.263	0.602	C15	0.120	−0.286	0.448	C24	−0.047	0.608	0.135
C8	0.167	0.646	C16	−0.300	0.449	0.188	C25	0.200	0.517	−0.035
			C17	−0.241	0.503	0.151
Contribution rate	55.55%	21.01%	Contribution rate	34.98%	25.91%	16.15%	Contribution rate	36.26%	30.95%	13.25%
Cumulative contribution rate	55.55%	76.56%	Cumulative contribution rate	34.98%	60.89%	77.04%	Cumulative contribution rate	36.26%	67.21%	80.46%

).

4.3. Results of the Comprehensive Evaluation

Based on the above indicator of load coefficients, evaluation indices can be calculated for the node system (X1), place system (X2), and industry system (X3). The weight coefficients of the three subsystems are calculated according to the hierarchical analysis to obtain a comprehensive evaluation index of the development potential (X) of the station area (Equation (6)).

X = 0.493 × 1 + 0.196 × 2 + 0.311 × 3

(6)

The comprehensive evaluation index of the development potential of the 18 station areas can be calculated according to Equation (6) (

Table 7. Development potential evaluation index and ranking of the HSR Stations.

Grade	HSR Station	Node Value	Node Ranking	Place Value	Place Ranking	Industry Value	Industry Ranking	Integrated Value	Integrated Ranking
Provincial Capital Hub Stations	Changshanan	1.538	1	0.700	1	0.735	1	1.129	1
Prefecture Hub Station	Huaihuanan	0.788	5	0.656	3	0.590	2	0.703	2
	Hengyangdong	0.836	3	0.578	4	0.282	11	0.616	4
Prefecture Station	Yueyangdong	0.862	2	0.664	2	0.390	5	0.679	3
	Chenzhouxi	0.832	4	0.352	11	0.384	6	0.601	5
	Zhuzhouxi	0.764	6	0.502	7	0.391	4	0.599	6
	Loudinan	0.740	7	0.498	8	0.338	8	0.570	7
	Xiangtanbei	0.642	9	0.283	14	0.313	9	0.471	9
County Station	Shaoshannan	0.707	8	0.565	5	0.301	10	0.555	8
	Leiyangxi	0.458	13	0.554	6	0.232	14	0.408	10
	Lilingdong	0.460	12	0.312	13	0.343	7	0.396	11
	Miluodong	0.414	14	0.420	9	0.257	12	0.368	12
	Zhijiang	0.548	10	0.386	10	0.048	18	0.362	13
	Xinhuangxi	0.476	11	0.339	12	0.089	16	0.330	14
	Shaoyangbei	0.236	15	0.002	18	0.557	3	0.291	15
	Xinhuanan	0.179	17	0.148	15	0.246	13	0.194	16
	Hengshanxi	0.185	16	0.031	16	0.066	17	0.118	17
	Xupunan	0.021	18	0.013	17	0.133	15	0.049	18

, Figure 7).

The study found that the development of transport space, physical space, and industrial space at the same station is mostly not synchronized and that there is a significant difference between the industrial value ranking and the ranking of each station area’s comprehensive potential assessment, with place value coming second and node value data being the closest and most reflective of each station area’s comprehensive potential level.

4.4. Results of the Hierarchical Clustering Analysis

Hierarchical cluster analysis in multivariate analysis under the Origin 2021 platform was carried out in the statistics module and the final results of the cluster analysis obtained through iteration and classification are shown in

Table 8. Results of the hierarchical clustering analysis: H, high degree; M, moderate degree; L, low degree.

Cluster Number	Broad Category	Sub-Category	Type	HSR Station
1	Dependent	NPI Dependent	L-L-L	Xinhuanan, Hengshanxi, Xupunan Railway Station
2	Class Dependent	Node Lightly Dominated Class Dependent	M-L-L	Lilingdong, Zhijiang, Miluodong, Xiangtanbei Railway Station
3	Class Dependent	Place Lightly Dominated Class Dependent	L-M-L	Xinhuangxi Railway Station
4	Balanced	NPI Balanced	M-M-M	Shaoshannan, Leiyangxi, Chenzhouxi Railway Station
5	Class Balanced	Node Deficient Class Balanced	L-M-M	Shaoyangbei Railway Station
6	Class Balanced	Industry Deficient Class Balanced	M-M-L	Loudinan Railway Station
7	Stressed	NPI Stressed	H-H-H	Huaihuanan, Changshanan Railway Station
8	Class Stressed	Industry Balanced Class Stressed	H-H-M	Yueyangdong, Zhuzhouxi, Hengyangdong Railway Station

. At the same time, the point distribution of the 18 site node, place, and industry values in a three-dimensional coordinate system was mapped in Origin 2021 using the 3D scatterplot function in plotting. The study used the 2D scatterplot function to plot the distribution of node values and place values, node values and industry values, and place values and industry values in a two-dimensional coordinate system to show the results of the cluster analysis more intuitively (Figure 8). Combined with the NPI model constructed in the previous section, the final station types in the NPI model corresponding to each station were obtained, for a total of eight categories.

Among them, Cluster 1 is the NPI Dependent type, which includes Xinhuanan Station, Hengshanxi Station, and Xupunan Station. The station’s low grade and remote location, low level of interdependence between node, place, and industry values, with small transport needs make it difficult to drive functional complex land development and business clustering in the station area. Cluster 2 belongs to Node Lightly Dominated Class Dependent, which contains Lilingdong Station, Zhijiang Station, Miluodong Station, and Xiangtanbei Station. The station does not follow the main direction of urban development and lacks development momentum at this stage, with delayed land development and low level of industrial development despite moderate transport values. Cluster 3 is the Place Lightly Dominated Class Dependent type, which includes Xinhuangxi Station, where the land development process is superior to transport organization and industrial clustering, with low node and industrial values and moderate place value. Cluster 4 is the NPI Balanced type, including Shaoshannan Station, Leiyangxi Station, and Chenzhouxi Station, with moderate node, place, and industry values that promote each other. Their land use, industrial agglomeration, and transport functions are compatible, with more reasonable station area development. Cluster 5 is the Node Deficient Class Balanced type, including Shaoyangbei Station, with moderate and mutually reinforcing place and industry values, but weak node function. The supply of traffic in the station area exceeds the demand, and the coordinated spatial development and rational functional positioning of the station area ensure a certain amount of industrial development. Cluster 6 is the Industry Deficient Class Balanced type, which includes Loudinan Station, where node and place are of moderate value and coordinated with each other relatively, however, the value of the industry is relatively low. Public transport provision in the station area matches the level of land development, however, with property development as the main focus, it is difficult to drive industrial development. Cluster 7 is the NPI Stressed type, including Huaihuanan Station and Changshanan Station, where node, place, and industrial development are saturated. The traffic carrying capacity is close to the upper limit, the land use potential is high and most of the land is already developed, and there is also a high concentration of industries. Cluster 8 is the Industry Balanced Class Stressed type, including Yueyangdong Station, Zhuzhouxi Station, and Hengyangdong Station, with high node and place values and relatively balanced industry values. The station area is close to saturation in terms of traffic demand and land development, while the distribution of industries is relatively concentrated. Overall, only Shaoshannan, Leiyangxi, and Chenzhouxi stations, as well as Shaoyangbei and Loudinan stations are in equilibrium or relatively equilibrium, with most stations in an uneven state.

In addition, there is a large variation in the sites found to be in unbalanced condition from the scatterplots of the three different 2D value combinations. The number of stations in unbalance in the node-place and node-industry scatterplots is six and seven, respectively, most of which are duplicated consistently. The 14 stations that are out of balance in the place-industry scatter plot indicate that the degree of mismatch between the place function value and the industry aggregation value in most station areas of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines is much higher than that of the other two two-dimensional value combinations, which also shows the importance of the industry aggregation value to the development of station areas.

5. Discussion

Multi-polarity exists in the development of China’s HSR station areas [52]. To assist decision makers in reasonably assessing the development status of HSR station areas, promoting sustainable development of station areas, and guiding different levels and types of HSR station areas to compete in a staggered manner, this paper introduces the industry dimension and constructs the node-place-industry model (NPI model) and innovates the evaluation index system and evaluation methods. We empirically analyzed the development potential of 18 HSR station areas along the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines and derived the development type of each station area through cluster analysis.

The results of the study show that, in general, the 18 stations along the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines can be classified into six major categories: Dependent, Class Dependent, Balanced, Class Balanced, Stressed, and Class Stressed, and eight minor categories: NPI Dependent, Node Lightly Dominated Class Dependent, Place Lightly Dominated Class Dependent, NPI Balanced, Node Deficient Class Balanced, Industry Deficient Class Balanced, NPI Stressed, and Industry Balanced Class Stressed. Similar to the findings of previous studies on the assessment of HSR station areas, most of the station areas of the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines are in a non-sustainable state of development. The new findings include the expansion of station area types from the five identified by the NP model to twenty-seven, and the main reason why most HSR station areas are in a state of disequilibrium is the mismatch between the functional value of place and the value of industrial aggregation.

The type of station is directly related to the location, the grade of the station, and the length of time the station area has been developed. In terms of station location, HSR station areas located on the edge of mega-cities and metropolitan areas are more likely to be developed as Stressed or Class Stressed, station areas located on the edge of urban areas in medium-sized cities and small towns are more likely to develop as Balanced or Class Balanced. Station areas located on the outskirts of the county or the edge of industrial parks are prone to develop as Class Dependent and station areas located in rural areas tend to develop into Dependent. In terms of station class, there is a clear positive trend in the integrated potential score and station class for most of the station areas along the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines. The higher the grade of the station, the better the accessibility, the stronger the function of the transport node, the higher the intensity of the land development carried, and the higher the density of the corresponding industrial gathering. That is, most of the Provincial Capital Hub Stations and Prefectural Hub Stations develop into Stressed and Class Stressed types, the general Prefecture Station prefers to develop into Class Balanced and Balanced types, a few into Class Stress types, and the county stations tend to develop into Dependent and Class Dependent types. In terms of development time of the HSR station area, the Hunan section of the Beijing–Guangzhou high-speed railway opened five years earlier than that of the Shanghai–Kunming high-speed railway. Four of the eight station areas on the Hunan section of the Beijing–Guangzhou high-speed railway are evaluated as Stressed or Class Stressed and two are developed as Balanced. Six of the ten station areas in the Hunan section of the Shanghai–Kunming high-speed railway are evaluated as Dependent or Class Dependent. Therefore, the longer the station area is developed and constructed, the more fully the node, place, and industry values are developed and the more likely it is to develop into a Balanced or Stressed type.

It is appropriate to adopt corresponding development strategies and methods for different types of station areas to promote the synergistic development of node, place, and industry spaces, to ensure that HSR station areas are conveniently connected with the urban centers and form a more inseparable spatial network through co-ordination and complementarity in terms of functions and industries. The NPI Dependent and Class Dependent station areas, often positioned as high-speed rail characteristic towns or high-speed rail clusters on the edge of cities, are mainly located in rural areas or on the edge of county towns and industrial parks, with poor transport infrastructure, low functional configurations, and limited levels of industrial development, unable to form a virtuous cycle of promotion. Shortly, transport and other ancillary facilities, and interaction with external areas, should be improved to enhance the design of transport connections with the town, urban areas, and industrial parks and to include amenities and facilities to bring in permanent people flow [2]. In the medium to long term, it can combine its resource endowment to cultivate cultural tourism, commerce and trade, and other special industries, as well as advanced manufacturing and productive service industries, and promote the development process of the station area through industrial clustering and attract regional arrival traffic, thus increasing the value of node and place and achieving sustainable development. NPI Balanced and Classed Balanced station areas are positioned mainly as an urban sub-center, with a few positioned as high-speed rail towns or urban fringe clusters. The development of related industries can be further driven by strategic engine projects. The selection and positioning of industries in the station area should emphatically reflect the regional and urban functional characteristics and be based on the existing basic industries to avoid homogenization of industries. with a balance between transport supply and demand and a relatively rich variety of land types. With a moderate concentration of industries, it helps to guarantee the stability of the land development process and type, as well as the sustainability of the traffic demand for arrivals within and outside the city, achieving a balanced development with positive feedback. For NPI Stressed and Class Stressed station areas, which are often positioned as urban sub-centers, they have a high degree of matching between land use, industrial agglomeration, and actual transport functions, and have good sustainability of development. The development of the station area will be strengthened by staggered development with the city center, focusing on fostering modern service industries or knowledge-intensive industries such as finance and commerce, logistics and trade, science and technology research and development, and culture and creativity for regional service, while shifting the focus of the construction of the station area to supporting public service facilities and environmental quality improvement and build an intellectual services infrastructure that promotes human interaction, knowledge creation and innovation [1]. To cope with the traffic congestion in the connected urban areas that may be caused by high node value, the spatial and temporal characteristics of congestion in different clusters of the transport network can be collected and a reliable traffic prediction system can be developed [53]. In particular, there is a need to avoid an overdevelopment of the station area, which could lead to a situation where the transport function of the station and the activation function of the area are in tension with each other beyond the maximum, which could ultimately deviate from sustainable development. The next step could be to combine the current status of the implementation of the efficiency, results, and benefits of each station area, explore the development mode and operation mechanism of HSR station areas, and propose targeted countermeasures and measures for balanced development, taking into account the actual situation of different types of station areas.

In contrast to other established relevant studies, this study is the first empirical study to construct an NPI expansion model to date and has the following features and innovations. Firstly, the assessment of the development potential of the HSR station area is linked to the three elements of the station area: transport space, physical space, and industrial space, emphasizing the key role of industrial aggregation in the development of station areas. Based on this, the industrial dimension is introduced and the NPI model is established. Secondly, a station typology framework that takes into account the three-dimensional elements of node, place, and industry has been established, defining 7 categories and 27 types of station area development, covering all station area types, which provide a counterpart for cluster analysis. Finally, this study is also innovative in terms of the evaluation indicator design and basic data acquisition. In terms of indicator design, not only have the industry dimension indicators been added, but the design of indicator items for the place dimension has also been optimized to form a more complete evaluation system. Evaluation indicators for the industry dimension and place dimension were constructed from the perspectives of industry–city integration and station–city integration, respectively. For the study data acquisition, the POI data of enterprises in 18 HSR station areas at the end of December 2022 were crawled to explore the possible application of new data and technology in the assessment of the development potential of HSR station areas. The introduction of the industry dimension not only expands the NP theoretical model methodologically but also expands the station area types from the 5 identified in the NP model to 27. Meanwhile, by revealing the key role of the industrial aggregation value, it provides a technical method for the empirical research of such typical transport hubs as high-speed railway stations, and the quantitative and cluster analysis results can be used to support the adjustment of the grading classification of various station areas.

6. Conclusions

The sustainability of the development of the HSR station area is the result of a combination of factors such as transport organization, land use, and industrial agglomeration. In this study, an NPI model was constructed to determine the magnitude of the sustainable potential of a station area by calculating a potential weighted score for the particular station area. By dividing the types and clarifying the characteristics of the current stage of development of stations, proposing countermeasures for the development of different types of station areas, and drawing the following three conclusions.

Firstly, the development of the three elements of the station area space is mostly unsynchronized, the node value has the highest fit with the comprehensive potential value of the station area, followed by the place value, and there is a large difference between the ranking of the industry value and the ranking of the comprehensive potential assessment of the station area. Secondly, there is a significant correlation between the overall potential of the HSR station area, the location and the grade of the station, and the length of development of the station area, and this directly affects the results of the HSR station type assessment. The closer the site is to the urban center, the higher the grade, and the longer the site has been developed, the higher the integrated potential ranking and the more likely it is to develop into a Stressed or Class Stressed type. Conversely, the lower the ranking, the more likely it is to develop into a Dependent or Class Dependent type. At the same time, the overall ranking of prefecture-level city station areas varies relatively little, indicating that the development potential of prefecture-level city station areas is relatively balanced. The overall ranking of county-level station areas varies greatly, with some having little development potential. Thirdly, based on the three-dimensional projection of the NPI model, 7 categories and 27 station area types have been refined. Based on the results of the cluster analysis, the 18 HSR stations along the Hunan section of the Beijing–Guangzhou and Shanghai–Kunming high-speed railway lines were classified into 8 corresponding types, and most of the stations were found to be in an unbalanced state.

Due to the relatively small sample size of cases in this study, only 8 of the 27 types of station areas were empirically studied, and the findings lack generalizability. At the same time, the corresponding development strategy is proposed only for the broad categories of stations, which may have a problem of poor targeting. In the future, a comprehensive and systematic study on 27 types of station areas can be conducted with more case samples. It will also explore the planning approaches and development strategies from “imbalance” to “balanced” development in the light of the current status of the development efficiency, results, and benefits of various types of station areas, providing an auxiliary decision-making basis for the planning and development of new station areas or the optimization, adjustment, and secondary development of existing station areas. In addition, the results of the potential evaluation are only quantitative measurements, the planning and construction of the station area involves the coordination and communication of all stakeholders and needs to pay attention to the needs of the users, operators, and managers at the same time to contribute to the sustainable development of the station area at a higher level with a human-centered approach. Further work may therefore include the construction of scenarios for the diversification of station area development types in response to the needs of different stakeholders, as well as the multi-criteria assessment of these scenarios.

In conclusion, despite the limitations and potential improvements discussed, this study contributes to improving the understanding of relevant researchers on the connotations of the potential of the HSR station area and expanding the ideas of the evaluation framework, providing an interesting assessment tool for the whole process of sustainable development of the transport hub area.