Evaluation and Spatial Characteristics of Cooperation among Tourist Attractions Based on a Geographic Information System: A Case Study of The Yangtze River Delta Region, China

Abstract

With the development of global economic integration and the gradual formation of unified tourism markets, strengthening regional tourism cooperation has become an internal requirement of regional tourism development but also a new way of sustainably developing tourism. This study selected the key factors affecting the cooperation of tourist attractions, including the competitiveness of tourist attractions and the relationships among tourist attractions, and established an evaluation index system and mathematical model of tourist attractions’ cooperation. Furthermore, the level of cooperation was evaluated. According to the value of the cooperation level, the spatial characteristics of the cooperation level were analyzed using a geographic information systems analytical method, which can better reflect the competitiveness, relationships, and overall cooperation level of tourist attractions. The results showed that the tourism competitiveness of tourist attractions was generally strong and their internal relations relatively close, and the overall tourism cooperation level was high. However, a two-dimensional four-quadrant map revealed that there were still great differences in tourism competitiveness among tourist attractions and their internal relations. Twenty-three tourist attractions exhibited weak tourism competitiveness and sparse relationships with other tourist attractions. The tourism competitiveness of tourist attractions and their internal relations and tourism cooperation level showed positive spatial autocorrelation and spatial agglomeration characteristics. The spatial differentiation of an “inverted U-shape” indicated that the cooperation level, tourism competitiveness, and mutual relations of tourist attractions were not balanced and that a stable and gradual spatial transformation had not been achieved. This study can provide valuable insights for the government to formulate policies and measures for regional tourism cooperation, carry out regional joint marketing, and help tourism enterprises design tourist routes.

1. Introduction

Tourism, today, has become an important subsystem of the global industrial system and, increasingly, a strategic means for tourist destinations to compete internationally [1]. Competition among tourism destinations presents some new characteristics and, especially with the gradual formation of the global tourism market, has led to the development of a unified international standard for tourism product quality [2]. The international tourism market promotes the standardization of tourism products, while only through cooperation and maximizing the needs of tourists can tourist destinations create their own advantages and compete for the limited tourism market [3].

Cooperation among tourism destinations refers to the organic cooperation between different tourism destinations based on the premise of competition [4,5,6,7,8,9,10]. Generally, when two or more cities choose to share resources in order to exploit developmental opportunities jointly, the rationale of cooperation is added to the one of competition [11]. Nguyen found that the competitive factor was necessary for the cooperative factor to work [12]. Under certain conditions and intervention, competition will eventually be transformed into cooperation [12]. From a biological perspective, Herring pointed out that it is an evolution from competition to cooperation in an ecosystem [13]. Yang believes that there will be a certain amount of competition before companies cooperate [14]. When enterprises realize that under the current competitive environment they cannot obtain the maximum potential profit, they will actively seek cooperation so as to achieve Pareto improvement [14]. Bolli et al. found that cooperation tendency depends on the competitive environment, and different dimensions of competition (i.e., the quantity, quality, and price competition of the main competitors) all motivate cooperation to a certain extent, which reflects that competition is the basis of cooperation [15]. From the perspective of business relations, Leite discussed that under the premise of competition, the higher the degree of interdependence, the more cooperation enterprises tend to have [16]. Hoffmann et al. also elaborated on the possibility of constraining versus reinforcing associations, whereby competition leads to cooperation between firms [17]. Hamel et al. addressed that collaboration is a competition in a different form, thus establishing competition as the starting point of cooperation [18]. One of the key points is how rival firms can coordinate themselves and construct collective actions without losing their individual competitiveness and flexibility [19]. Therefore, we believe that tourist attractions should be in a competitive environment before cooperation, and they should have a certain level of competitiveness [20]. In essence, tourism destinations possess certain attractions and competitiveness through their own development, and in order to seek greater competitive advantages, they, along with other destinations, jointly launch a stable, harmonious, mutually beneficial, and dynamic balance of space cooperation [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23].

Regarding the key factors affecting cooperation between tourism destinations, scholars have put forward their own views from different perspectives. For example, Teye believed that the following factors directly determine the success or failure of cooperation between tourist destinations: diversified tourism destination development, joint promotion and marketing, tourism facilitation policies (including liberalization of passport and visa requirements), transport system policies, standardization and regional classification of tourism facilities, and conservation and preservation programs for natural, historical, and cultural resources [24]. Wang et al. argued that the forces of competition, economic conditions, technology, and risk lead potential cooperative partners to solve common problems [25]. Fyall and Carrod identified international political and trade agreements, capital accumulation and concentration, technological progress, and strategic alliances as key factors influencing cooperation in tourism destinations [26]. Li et al. classified the various influencing factors of regional tourism cooperation into exogenous, endogenous, and random factors, mainly including economic, political, sociocultural, legal, demographic, and geospatial factors [27]. Yin et al. identified a set of factors that affect tourism cooperation in China–ASEAN including differences in political system, governance, and income [28]. In addition to the abovementioned factors influencing cooperation between tourism destinations from the perspective of supply (including governments and enterprises), some scholars have also analyzed factors influencing cooperation between tourism destinations from the perspective of demand [29,30,31,32]. The multidestination tourism behavior of tourists is regarded as the driving factor of cooperation among tourism destinations [29,30,31,32]. The spatial behavior and movement of tourists’ free choice objectively connect different tourist destinations. Therefore, these tourist destinations can establish mutual trust, respect, and other interactive relations through communication, thus forming a mutually beneficial cooperative relationship [29,30,31,32].

Previous studies on the evaluation of tourism destination cooperation have mainly been based on the key factors needed to enhance the partnership. For example, according to Gray and Wood, most theories focus on evaluating the final outcome of cooperation in terms of partnerships [33]. Katarzyna analyzed the partnerships among members of a destination marketing organization operating in southern Poland and evaluated the level of cooperation [1]. Yang claimed that the cooperation of tourist attractions refers to the interaction or collaboration between two or more tourist attractions in order to share information and tourist markets, jointly innovate tourism products, and achieve regional tourism development goals [29]. Thus, Yang evaluated the degree of cooperation between tourist attractions based on how often their information appeared and co-appeared in travel notes and official news items [29]. Although researchers have clarified the key factors affecting cooperation between tourism destinations from different perspectives, these factors also partly reflect the competitiveness of tourism destinations themselves and their cooperation with other tourism destinations and are basically sporadic and fragmented. Prior studies have failed to systematically examine the influencing factors of tourism destination cooperation from the perspective of a tourist destination’s own attractiveness (competitiveness) and its relationship with other tourist destinations as well as to evaluate cooperation based on these factors. In addition, most of the relevant studies on the cooperation evaluation of tourism destinations are based on the relationship formed among tourism destinations [30,31,32,33,34,35], and there is a lack of a comprehensive evaluation of cooperation among tourism destinations based on the above factors. Empirical research on the cooperation evaluation and spatial differentiation of tourist attractions in the Yangtze River Delta region of China is even rarer. In order to fill this research gap, this study used multisource data, including official statistical data, geospatial data, and transaction and tourism route data of online travel agency (OTA) websites, to evaluate the cooperation among tourist attractions in the Yangtze River Delta region of China and to analyze its spatial characteristics. Consequently, in this study, we address the following research questions: What are the key factors affecting the cooperation of tourist attractions in the Yangtze River Delta region of China? What are the factors that reflect the competitiveness of tourist attractions and the relationships among these factors? How can the cooperation of tourist attractions be evaluated based on these factors? Is there a spatial difference in the cooperation level of tourist attractions? If there are spatial differences, what are their spatial characteristics? These questions reflect the evaluation index system for cooperation among tourism destinations and the evaluation and level of cooperation between tourism destinations and their spatial characteristics. They can also help the government formulate policies and measures of regional tourism cooperation and tourism enterprises, tourism planning, and tourism circuit design and to carry out regional joint marketing to provide valuable insights. To answer these questions, we undertook the following two tasks.

First, based on previous studies and the connotation of cooperation in tourist attractions, we selected the key factors affecting cooperation among tourist attractions, including two indicators reflecting the competitiveness of tourist attractions and relationships among tourist attractions, and established an evaluation index system for cooperation among tourist attractions. We adopted the entropy weight method to determine the index weight according to the index differentiation degree. In addition, we created the cooperation evaluation model of tourist attractions and evaluated their cooperation level. Second, we further analyzed the spatial characteristics of the cooperation level of tourist attractions by using a geographic information systems (GIS) analytical method, which can better reflect the competitiveness, cooperative relations, and overall cooperation level of tourist attractions.

2. Materials and Methods

2.1. Study Area

Fifty-five 4A and 5A tourist attractions in 16 cities in the Yangtze River Delta region were selected as the research samples. In China, 5A tourist attractions are rated by the People’s Republic of China based on the level of quality of the tourist attraction. The level of quality is divided into five levels: 5A, 4A, 3A, 2A, and A in descending order. 5A tourist attractions in China are rated by the Quality Rating Committee of Tourist Attractions set-up by the Ministry of Culture and Tourism, PRC, from 4A tourist attractions, and then the quality rating of tourist attractions is produced. Both 5A tourist attractions and 4A tourist attractions are high-level tourist attractions in China, representing the level of world-class, high-quality tourist attractions in China. The 16 cities include Shanghai, Nanjing, Suzhou, Yangzhou, Zhenjiang, and Taizhou, in Jiangsu Province, and Wuxi, Changzhou, Nantong, Hangzhou, Ningbo, Huzhou, Jiaxing, Zhoushan, Shaoxing, and Taizhou in Zhejiang Province (Figure 1). According to the Regional Planning for the Yangtze River Delta region (2008–2010), these 16 cities are the most representative core cities in the Yangtze River Delta region. The 4A and 5A tourist attractions were selected according to the “List of China’s A-level Tourist Attractions”. On the one hand, 4A and 5A tourist attractions are selected for research because of their attractiveness and competitiveness [36]. On the other hand, tourists tend to visit high-level tourist attractions [37,38,39]. The above two reasons make it possible to form a relatively close cooperative relationship among these tourist attractions.

2.2. Research Framework for Analyzing the Cooperation among Tourist Attractions

A new research framework was designed in this study, as shown in Figure 2. This study used multisource data, including official statistical data, geospatial data, and transaction data and tourism route data of an OTA website, to evaluate the cooperation among tourist attractions in the Yangtze River Delta region of China. First, this study selected four indexes of attraction of tourism resources, instantaneous bearing capacity, number of visitors, and tourism income to analyze the competitiveness of tourist attractions and selected two indexes of connectivity and co-occurrence frequency to analyze the relationships among tourist attractions. Second, on this basis, the cooperation of tourist attractions was evaluated. In addition, with CITA as the x-axis, RITA as the y-axis, and the average value of CITA and RITA as the origin, a plane rectangular coordinate system was constructed according to the scores of CITA and RITA, and the competitiveness scores of tourist attractions and their relationship with other tourist attractions were compared and analyzed. Third, the spatial pattern of the cooperation levels of the tourist attractions was analyzed. In order to observe the spatial pattern of the cooperation levels of the tourist attractions more intuitively, using Moran’s I indexes, the kernel density analysis in the GIS analysis method, spatial trend analysis, and hotspot analysis showed spatial visualization and comparison of cooperation among tourist attractions.

2.3. Cooperation Evaluation of Tourist Attractions (CETA)

Referring to the existing research, this study constructed the evaluation index system of cooperation between tourist attractions using the two aspects of competitiveness and relationship [24,25,26,27,28,29,30,31,32]. The competitiveness of tourist attractions is expressed by four indexes: tourist resources, tourist capacity, number of visitors, and tourist income. The relationship between tourist attractions is represented by two indicators: the connectivity index and co-occurrence word frequency.

2.3.1. Competitiveness Index of Tourist Attractions (CITA)

Attraction of Tourism Resources: First, the attraction of tourism resources (ATR) is a key index to measure the competitiveness of tourist attractions [36]. The classification of China’s A-level tourist attractions is based on the attractiveness of tourism resources as a prerequisite. Therefore, the different levels of A-level tourist attractions can be regarded as an important standard to measure the attractiveness of tourism resources. Second, the ATR also needs to be reflected by the attention of potential tourists. This study chose the Baidu Index to reflect the degree of attention of tourist attractions. The Baidu Index, as a data statistical analysis platform, is of great value for exploring users’ needs. The Baidu Index can reflect users’ online attention to a certain keyword, and it is expressed through internet channels with specific numbers such as clicks and searches. After entering keywords in the Baidu Index platform, trend research, demand map of search heat and distribution of demand groups will be presented in different sections within a period of time. Researchers can type keywords into web pages to get their Baidu Index. Hence, in order to obtain the national tourist search trends and the online attention of tourists to 55 tourist attractions, the name and abbreviation of the tourist attractions in this study were entered in the official website of the Baidu Index.

The steps of data collection and grading were as follows: First, we went through the official government websites of the 16 cities to collect the latest list of A-level tourist attractions, from which we selected a total of 55 AAAAA and AAAA tourist attractions. We assigned 10 and 8 points to AAAAA- and AAAA-level tourist attractions, respectively. Second, we entered the names of the 55 tourist attractions in turn on the Baidu Index’s official website (https://index.baidu.com/v2/index.html#/, 15 December 2021). Then, the overall daily average of the Baidu Index of all the tourist attractions for all time periods (up to 15 December 2021) was obtained. We organized the Baidu Indexes of the tourist attractions to determine their attention. A Baidu Index of above 1000 is extremely high and was thus assigned a unified value of 10 points. Tourist attractions that had a Baidu Index of below 1000 were observed to have obvious breaks every 100 and, thus, 100 was used as the interval for division. A Baidu Index between 900 and 1000 was assigned nine points, and that between 800 and 899 was assigned eight points. In this way, a Baidu Index of less than 100 was assigned one point, which is extremely low. The ATR is expressed as follows:

$$AT{R}_i=LT{A}_i+D{A}_i$$

(1)

where ATR_i represents the tourism resource attraction of the i-th tourist attraction; LTA_i represents the grade score of the i-th tourist attraction; DA_i indicates the attention score of the i-th tourist attraction. The sum of LTA_i and DA_i is the attraction of tourism resources of the i-th tourist attraction.

Instantaneous Bearing Capacity: Selecting an instantaneous bearing capacity (IBC) index is of great significance for maintaining the competitiveness of tourist attractions and for promoting the sustainable development of tourist attractions [39]. IBC refers to the maximum number of tourists that a tourist attraction can accommodate at a certain point in time under the premise of ensuring the safety of tourists in each tourist attraction and the safety of tourism resources and the environment. IBC is a comprehensive capacity including ecological capacity, environmental capacity, facility capacity, and social capacity, which can measure the size of tourist space and tourist stay time. The larger the land area, the more tourists it can accommodate, and the greater the IBC of the tourist attraction’s resources. IBC is expressed as follows:

$$IB{C}_i=\sum \frac{X_i}{Y_i}$$

(2)

where IBC_i represents the instantaneous bearing capacity of tourist attraction i; X_i is the effective tourist area of tourist attraction i; Y_i is the unit area of tourists in tourist attraction i, namely, the basic space bearing standard. The data of X_i and Y_i come from a document (2015) on the maximum bearing capacity of 5A-level tourist attractions released by China’s Ministry of Culture and Tourism.

Number of Visitors: The number of visitors (NV) is the best indicator to reflect the competitiveness of tourist attractions [40]. Based on the transaction data of an OTA website, the stripping factors of tourist attractions in the tourism market of the counties (urban areas) where they are located were determined. A stripping factor is the ratio of the number of online tourism transactions in tourist attractions to the sum of the number of online tourism transactions in all tourist attractions in the same county (urban area). Then, the regional county (urban area) scale of tourism statistics was studied to calculate the tourist reception scale. NV is expressed as follows:

$$N{V}_i=S{C}_{i}TR{N}_i; S{C}_i=\frac{NO{T}_i}{{{\displaystyle \sum }}_i^{n}NO{T}_i}$$

(3)

where NV_i represents the number of visitors in tourist attraction i; NOT_i represents the number of online tourism transactions in tourist attraction i (i.e., the sum of tourist transaction volume when tourist attraction i appears on the route); SC_i represents the stripping coefficient factor of tourist attraction i; TRN_i refers to the annual tourist reception number of the county (urban area) where tourist attraction i is located, and it can be obtained through the statistical yearbook; N represents the number of tourist attractions in the same county (urban area) as tourist attraction i.

Tourism Income: Tourism income (TI) is another important indicator that reflects the competitiveness of tourist attractions [41]. Assuming that the per capita consumption level of county (or urban area) tourist attractions is basically the same, TI is the product of per capita consumption level and NV (obtained above). TI is expressed as follows:

$$T{I}_i=PC{C}_{i}N{V}_i; PC{C}_i=\frac{AT{I}_i}{A{T}_i}$$

(4)

where TIi represents the TI of tourist attraction i; PCCi refers to the per capita consumption level of the county (or urban area) where tourist attraction i is located; AT_i and ATI_i, respectively, represent the annual tourist reception number and annual TI of the county (or urban area) where tourist attraction i is located. In addition, ATI_i can be obtained through the statistical yearbook.

2.3.2. Relationship Index of Tourist Attractions (RITA)

Connectivity Index: The establishment of cooperative relations is closely related to the traffic conditions of tourist attractions [42]. If the ability of tourists to travel to preferred destinations is inhibited by inefficiencies in the transport system, such as a far away or lengthy journey, the likelihood that they will seek alternative destinations may increase [42]. Hence, as a non-negligible factor affecting tourist flow, the traffic connectivity of tourist destinations should be regarded as one of the important characteristics of cooperative relationship attributes of tourist attractions. The connectivity index (CI) reflects the convenience of transportation between two tourist attractions, taking into full consideration the physical distance between tourist attractions with cooperative relationships and the normal time consumed by various road transportation modes. The transportation mode with the shortest time consumed between tourist attractions was selected for calculation and measured in minutes. The CI was introduced. Considering the excessively small value obtained by reciprocal processing, the calculation of connectivity was based on the rounding of the highest arrival time of the two farthest tourist attractions in the region, with the actual traffic time between cooperative tourist attractions subtracted. Finally, the CI was obtained by logarithmic processing. The data came from amAP. CI is expressed as follows:

$$C{I}_{ij}=\ln \left(T_{The longest time }-T_{The actual time}\right)$$

(5)

Co-Occurrence Frequency: If two tourist attractions appear at the same time on a tourist route, we recorded them as appearing simultaneously [29]. The sum of the co-occurrence of two tourist attractions is the sum of the co-occurrence frequency (CF) of tourist attractions on a tourist route. The frequency of co-occurrence words reflects the connection between tourist attractions and can be used as an important indicator of cooperation between tourist attractions. Transaction data from online travel platforms were used as data sources to calculate the occurrence frequency of co-occurrence words. The data were dichotomized, where 0 indicates no connection and 1 indicates connection. By symmetrizing the cooperation matrix of the tourist attractions, the co-occurrence matrix was obtained, and the sum of the ranks and columns of tourist attraction i in the matrix is the CF of tourist attraction i.

2.4. Research Methods

2.4.1. Construction of the Mathematical Model

Cooperation among tourist attractions includes the competitiveness of tourist attractions and the relationship between them. According to the connotation and significance of the selected indicators, ATR, IBC, NV, and TI belong to the competitiveness of tourist attractions, while CI and CF belong to the relationship formed between tourist attractions. The entropy weight method was adopted to construct the evaluation model of cooperation among tourist attractions (CTA) as follows:

$$CIT{A}_i=w_{11}AT{R}_i+w_{12}IB{C}_i+w_{13}N{V}_i+w_{14}T{I}_i$$

(6)

$$RIT{A}_i=w_{21}C{I}_i+w_{22}C{F}_i$$

(7)

$$CET{A}_i=w_{1}CIT{A}_i+w_{2}RIT{A}_i$$

(8)

where CETA_i, CITA_i, and RITA_i represent the cooperation level, competitiveness, and cooperation of the i-th tourist attraction, respectively. To eliminate dimensional effects, all variables were normalized. w represents the weights of indicators at all levels.

2.4.2. Spatial Analysis

Based on the geographic database of the cooperation among tourist attractions, the spatial correlation and heterogeneity of the cooperation level of tourist attractions were revealed with the help of ArcGIS 10.5 software and spatial statistical techniques, such as kernel density, hot spots, and spatial autocorrelation, and the spatial differentiation characteristics of the cooperation level of tourist attractions were visualized and analyzed.

2.5. Data Sources and Processing

2.5.1. Data Sources

This study used multisource data, including official statistical data, geospatial data, and transaction data and tourism route data of OTA websites, to evaluate the cooperation among tourist attractions in the Yangtze River Delta region of China.

Number of Visitors in Tourist Attractions: Based on the transaction data of OTA websites, the stripping factors of tourist attractions in the tourism market of their counties (urban areas) where they are located were determined. We obtained the complete itinerary data and accumulated transaction data of tourist attractions from the websites of Ctrip (www.ctrip.com), Tongcheng (www.ly.com), and Tuniu (www.tuniu.com) (accessed on 8 October 2021), which were released by tourism enterprises in group tour and independent tour products. To calculate the number of tourists received by tourist attractions, it was assumed that the actual flow of tourists will depend on the travel route. Ctrip, Tongcheng, and Tuniu are the top three one-stop travel platforms in China. Their websites provide free information regarding accommodation reservations, transportation tickets and itineraries issued by travel companies, and tourists’ travel diaries. All this information is publicly and freely available.

Shortest Transportation Time: The actual distance between each tourist attraction and the speed of the most convenient mode of transportation were obtained through amAP big data.

Other Sources of Data: Other data, such as tourist attention, were obtained by calculating the overall daily average value of the Baidu Index of all tourist attractions in all time periods by inputting the tourist attraction’s name and its abbreviation in the official website of the Baidu Index. In addition, we obtained the grade, floor area, tourism income, and other data on the tourist attractions through the official website of the local government.

2.5.2. Data Processing

Selection of Tourist Attractions: To select tourist attractions, we, first, undertook data cleansing. For repeated travel routes, specifically, we only retained one tourist attraction and removed the other redundant ones. Tourist routes not within the defined study area were also eliminated. Second, we extracted tourist attractions one by one from the retained tourist routes. On the premise of merging multiple tourist attractions belonging to the same scenic area (for example, Wuzhen Dongzha and Wuzhen Xizha belong to “Wuzhen”), tourist attractions were extracted. Consequently, we collected a total of 26,061 tourist routes and screened 16,666 effective tourist routes. These effective tourist routes contained 383 tourist attractions in total, from which we extracted 55 effective tourist attractions.

Data Standardization Processing: Multisource data are limited by dimension inconsistency and horizontal comparability; therefore, in this study, we adopted standardization to solve these problems. Normalization of data is scaling the data so that they fall into a small, specific range. It is often used in some comparison and evaluation index processes to remove the unit limit of data and transform it into a dimensionless pure value, so that indicators of different units or levels can be compared and weighted. The most typical one is the normalization of data, that is, data is mapped uniformly to the interval [0,1]. Min–Max normalization, also known as deviation normalization, is a linear transformation of the original data that results in the range [0,1], where Max is the maximum value of the sample data and Min is the minimum value of the sample data.

Index Weight: In this study, the entropy weight method was used to determine the weight. First, a two-module data matrix was constructed from the observed values of the objects to be evaluated on each index. After dimensionless standardization, the proportion of the index value of the object to be evaluated under each index was calculated to form a judgment matrix. Then, the entropy value and entropy weight of each index was calculated, and finally, the comprehensive weight of the index, namely, the weight, was obtained. The standard values of the observation data of 55 tourist attractions to be evaluated under eight indicators were taken as the original data matrix, and the operational process of the entropy weight method was strictly implemented. Finally, the weights of the cooperation evaluation indicators of the tourist attractions were obtained (

Table 1. Weights of the cooperation evaluation indexes of the tourist attractions.

Serial Number	Index	Weight Code	Weight Value
1	Competitiveness index	W1	0.6643
1-1	Attraction of tourism resources	W11	0.2892
1-2	Instantaneous bearing capacity	W12	0.0781
1-3	Number of visitors	W13	0.1416
1-4	Tourism income	W14	0.1554
2	Relationship index	W2	0.138
2-1	Connectivity index	W21	0.1211
2-2	Co-occurrence frequency	W22	0.0169

).

3. Results

3.1. Evaluation of the Cooperation among Tourist Attractions

According to Equation (8), the CETA, CITA, and RITA scores of the 55 tourist attractions were calculated and ranked (

Table 2. Cooperation index and subindexes.

Code	Name of the Tourist Attraction	City	CITA		RITA		CETA
			Score	Rank	Score	Rank	Score	Rank
N18	Wuzhen	Jiaxing	0.4319	3	0.6706	3	0.5120	1
N16	West Lake	Hangzhou	0.3665	11	0.7216	2	0.4857	2
N37	Classical Gardens of Suzhou	Suzhou	0.3681	10	0.6146	5	0.4508	3
N1	Qing He Lane	Hangzhou	0.2487	21	0.7977	1	0.4330	4
N35	Lingshan Buddhist Scenic Spot	Wuxi	0.4245	4	0.3303	13	0.3929	5
N24	Fuzimiao (Confucius Temple) Qinhuai Scenic Area	Nanjing	0.3577	13	0.4560	7	0.3907	6
N15	Songcheng	Hangzhou	0.2765	19	0.5865	6	0.3805	7
N27	Nanjing Zhongshan Mountain National Park	Nanjing	0.3281	17	0.3868	12	0.3478	8
N13	Nanxun Ancient Town	Huzhou	0.4682	1	0.0993	25	0.3444	9
N12	QianDao Lake	Hangzhou	0.4669	2	0.0886	28	0.3399	10

CITA, competitiveness index of tourist attractions; RITA, relationship index of tourist attractions; CETA, cooperation evaluation of tourist attractions.

). Due to the space limitation, this paper only presents the partial results of the cooperation index and subindexes.

The average value of RITA, CETA, and CITA was 0.1755, 0.1665 and 0.1933, respectively. There was no significant difference between the three scores. However, by comparing extreme values, it was found that Wuzhen had the highest RITA, which was 30.8 times higher than that of CCTV Wuxi Studios. The extreme value multiplier of CITA and CETA was 67.8 and 41.6 times different. Item comparison revealed 30 cooperative units with a difference of more than 10, accounting for 54.55% of the total.

With CITA as the x-axis, RITA as the y-axis, and the average values of CITA and RITA as the origin, a plane rectangular coordinate system was constructed according to the CITA and RITA scores (Figure 3). Taking the origin as the boundary, the value of CITA determines the competitiveness of the tourist attraction, while the value of RITA determines the density of the relationship between the tourist attractions. Different combinations of CITA and RITA scores can be assigned to the four quadrants of the coordinate system.

Wuzhen and the nine other tourist attractions located in the first quadrant had obvious advantages in competitiveness and relationships with other tourist attractions, which were distributed in Zhejiang and Jiangsu. This was mainly because these tourist attractions have unique tourism resources, with strong tourism attraction. In addition, these tourist attractions have better transportation facilities such that capital and tourism flow are concentrated here. Therefore, the relationship between these and other tourist attractions was also very close.

The second quadrant included nine tourist attractions located in cities such as Shanghai, Nanjing, and Hangzhou. These cities have achieved long-term sustained and rapid social and economic development, with obvious geographical and transportation advantages and relatively perfect cooperation conditions. However, compared with the cities in the first quadrant, the number of AAAAA tourist attractions in these cities was fewer, and the tourist attraction was slightly insufficient. In addition, due to the influence of theme culture, some tourist attractions (such as Disneyland) attract younger tourists and have limited source groups, resulting in low competitiveness of the tourist attractions.

The third quadrant comprised 23 tourist attractions such as the Shanghai Science and Technology Museum. These tourist attractions were weak in competitiveness, attracting little attention from tourists. They also had sparse relations with other tourist attractions. These tourist attractions were all developing tourist attractions and are expected to evolve and move into the second or fourth quadrant through policy guidance.

The fourth quadrant included 14 tourist attractions such as Nanxun Ancient Town. These tourist attractions had obvious competitive advantages, but their connection with other tourist attractions was slightly insufficient compared with the tourist attractions in the first quadrant. For example, Nanxun Ancient Town and QianDao Lake possess relatively high-quality tourism resources; however, there were relatively few connections between them and other tourist attractions due to their long distance and general traffic conditions.

3.2. Spatial Characteristic Analysis

A variety of spatial analysis methods were used in this study including spatial autocorrelation analysis, kernel density analysis, spatial trend analysis, and spatial interpolation and hotspot analysis. Among them, Moran’s I index is a comprehensive evaluation index used to measure the degree of spatial autocorrelation. However, Moran’s I index does not indicate where spatial aggregates of related elements exist and how they vary in space. Therefore, kernel density analysis, spatial trend analysis, and hotspot analysis were further carried out. The purpose was to verify the Moran’s I index to further explore the spatial aggregation phenomenon of relevant elements in specific locations through kernel density analysis and hot spot analysis and help understand the spatial change trend of relevant elements through spatial trend analysis. This can also be viewed as a continuous line of analysis for spatial analysis.

3.2.1. Spatial Autocorrelation Analysis

ArcGIS 10.5 software was used to analyze the spatial autocorrelation. As shown in

Table 3. Spatial correlation analysis of cooperation among tourist attractions.

Item	Global Autocorrelation			Local Autocorrelation
	Moran’s I	Z-Score	p-Value	Local G-Coefficient	Z-Score	p-Value
CITA	0.360946	4.755891	0.000002	0.023128	2.491002	0.012738
RITA	0.167501	2.380184	0.017304	0.025017	2.984260	0.002843
CETA	0.204388	2.809549	0.004961	0.020751	1.599265	0.109762

CITA, competitiveness index of tourist attractions; RITA, relationship index of tourist attractions; CETA, cooperation evaluation of tourist attractions.

and Figure 4, the Moran’s I indexes of CETA, CITA, and RITA were all greater than 0, showing a positive spatial autocorrelation and a certain spatial agglomeration. The z-scores of both CITA and CETA were greater than 2.58, and the p-value was less than 0.01, indicating significant agglomeration characteristics, which to some extent reflect the formation of a common and good cooperative relationship between tourist attractions, resource sharing, and reciprocal development. However, the z-score of RITA (2.380184) was greater than 1.96, and the p-value was less than 0.05, indicating significant agglomeration characteristics of spatial distribution, which may be related to the advantage of spatial proximity. The local G-coefficients of both CITA and RITA were greater than 0, their z-scores were greater than 1.96, and the p-value was less than 0.05. Both the local G-coefficients and z-scores were significant, indicating that local areas with high values tend to cluster and present hotspots. It also shows that tourist attractions with strong tourism competitiveness and close relationships tend to be adjacent, which is more conducive to resource sharing and win-win cooperation.

3.2.2. Kernel Density Analysis

Kernel density can be used to represent the spatial aggregation degree of the CETA, CITA, and RITA indexes. The higher the kernel density, the stronger the agglomeration effect and vice versa. Figure 5 shows the local aggregation of the CETA, CITA, and RITA indexes, which further verifies the results of the spatial autocorrelation analysis. According to the kernel density analysis based on the RITA index, two high-density core groups, Shanghai Oriental Pearl Radio and TV Tower and Qing He Lane, were formed in space. The closely related tourist attractions gradually spread from the high-density core group in Shanghai and Hangzhou to other tourist attractions in the province (city). Based on the kernel density analysis of CITA, a high-density core group of Tongli Ancient Town was formed in space. The tourist attractions with strong tourism competitiveness gradually spread from the high-density core group of Suzhou to other tourist attractions in the city and the province. The kernel density analysis based on CETA reflected the level of cooperation among tourist attractions, and the result shows that Shanghai Oriental Pearl Radio and TV Tower, Qing He Lane, Tongli Ancient Town, and Fuzimiao (Confucius Temple) Qinhuai Scenic Area constitute the four core groups with a high density and high level of cooperation of tourism scenic areas, with the high-density core of Shanghai, Hangzhou, Suzhou, and Nanjing as the center, gradually spreading to other tourist attractions inside the province (city).

3.2.3. Spatial Trend Analysis

Taking the changes in the longitude of CETA, RITA, and CITA in the “east–west” direction as the x-axis, the changes in the latitude of CETA, RITA, and CITA in the “south–north” direction as the y-axis, and the sizes of CETA, RITA, and CITA as the z-axis, a three-dimensional trend surface analysis was carried out (Figure 6). In the east–west direction, CITA and CETA showed an inverted U-shaped spatial differentiation, indicating that the cooperation level and tourism competitiveness of tourist attractions were lower in the east and in the west and higher in the middle. In the north–south direction, CETA and RITA showed an inverted U-shaped spatial differentiation, that is, the cooperation level of tourist attractions and the relationship among tourist attractions were lower in the north and in the south and higher in the middle. The spatial differentiation of an “inverted U-shape” indicates that the cooperation level, competitiveness, and mutual relations of the tourist attractions have not achieved stable and gradual transformation, either from the east–west direction or from the north–south direction.

3.2.4. Spatial Interpolation and Hotspot Analysis

Based on the cooperation index of tourism tourist attractions, spatial interpolation and hotspot analysis were used to classify them into hotspot regions, subhotspot regions, weak hotspot regions and general hotspot regions. The analysis found that the hotspot areas of CETA, RITA, and CITA showed an obvious concentration trend, which again verified the result of local autocorrelation analysis (Figure 7). No cold spot areas were found, mainly because AAAA and AAAAA tourist attractions with a good cooperation level were selected in this study, thus, resulting in no cold spot areas.

By comparing the hotspots of CETA, RITA, and CITA, it was found that the hotspots of CETA and RITA were generally the same, which were all in the vicinity of Shanghai, Nanjing, Hangzhou, Suzhou, Huzhou, Jiaxing, and Wuxi. These cities showed high cooperation level and strong tourism competitiveness. High-speed rail, aviation, and other rapid transport services bring about the spatiotemporal compression effect, greatly enhance the connection of tourist attractions, and improve the frequency of tourists’ visits, thus enhancing the spatial aggregation of the cooperation level index and relationship index of tourist attractions along the line. Nanjing, Hangzhou, Ningbo, Shaoxing, Suzhou, Huzhou, Jiaxing, and Wuxi had high values of CITA agglomeration, indicating that the tourist attractions in these cities had high tourism resource endowment and strong tourist attraction and were favored and concerned with by tourists. A few secondary and weak hotspots were scattered in the periphery of CETA, RITA, and CITA, and the distribution of weak hotspots was obviously different. In general, Zhoushan, Ningbo, Shaoxing, and southwest Hangzhou in the southern region had relatively high tourism resource endowment and attracted high attention from tourists, but they did not have superior cooperation conditions and had relatively few connections. Yangzhou in the north had average CETA, RITA, and CITA values, which need to be further improved. Compared with the central region, Zhoushan in the southwest and east of Hangzhou had average CETA levels. Shanghai, Suzhou, Wuxi, northeast of Hangzhou, Nanjing, and Jiaxing in the central part of the Yangtze River Delta had high CETA, RITA, and CITA values and were densely distributed, which further verifies the trend surface analysis results.

4. Discussion

Although the cooperation among tourist destinations is of great significance, the understanding of the cooperation among tourist attractions in the Yangtze River Delta is still limited. In order to make up for the deficiency that the existing literature usually only considers the cooperation relationship between tourist attractions [29,30], this study systematically investigated the influencing factors of the cooperation of tourist attractions from two aspects (i.e., the competitiveness of tourist attractions and their relationship with other tourist attractions) and evaluated the cooperation of tourist attractions in the Yangtze River Delta based on these factors. This study further analyzed its spatial structure characteristics using a GIS analytical method, which can better reflect the competitiveness, cooperative relations, and overall cooperation level of tourist attractions.

Among the top ten tourist attractions in terms of cooperation level, there are seven historical and cultural tourist attractions (i.e., Wuzhen, Classical Gardens of Suzhou, Qing He Lane, Lingshan Buddhist Scenic Spot, Fuzimiao (Confucius Temple) Qinhuai Scenic Area, Nanjing Zhongshan Mountain National Park, and Nanxun Ancient Town), two natural ecological tourist attractions (i.e., West Lake and QianDao Lake), and one modern amusement tourist attraction (i.e., Songcheng). This shows that historical and cultural tourist attractions are either highly competitive, closely connected with other tourist attractions, or both, so there are more tourist attractions with a high level of cooperation. In addition, four of these tourist attractions are located in Hangzhou, two in Nanjing, and one each in Jiaxing, Suzhou, Wuxi, and Huzhou. This was mainly due to the high quality of tourism resources and good traffic accessibility in these cities, which has a positive role in promoting the cooperation of tourist attractions [41,42,43,44,45,46]. This is quite different from the research conclusions of Wang et al. It is because this study included both competitiveness and relationship indicators in the evaluation of cooperation among tourist attractions, instead of only considering relationship indicators like Wang et al. As a result, the top ten tourist attractions obtained in this study are quite different from their own.

According to the constructed two-dimensional four-quadrant graph (Figure 1), we found that there were differences in the development of the competitiveness of different tourist attractions and their relationship with other tourist attractions. There were only nine tourist attractions with good development in both aspects. The nine tourist attractions are mainly located in Zhejiang and Jiangsu provinces. It can be seen that although municipal governments in the Yangtze River Delta have continuously held the “Yangtze River Delta Tourism City Summit Forum” since 2003, signed regional tourism cooperation agreements, and advocated for the construction of barrier-free tourism areas, an increasing number of cities have participated in the cooperation. But the effect of cooperation between tourist attractions is not satisfactory. The theoretical basis for this finding may be that most cooperation among tourist attractions is driven by government or business intervention rather than market mechanisms, or the market mechanism does not play a leading role in the cooperation of tourist attractions, leading to the unbalanced cooperation of tourist attractions. This supports the views of Yin et al. [28].

In terms of the spatial agglomeration characteristics of tourist attraction cooperation, the high values of CETA were mainly concentrated in Shanghai, Nanjing, and Hangzhou and gradually decreased in the surrounding areas. The hot spots of CETA, RITA, and CITA values showed spatial agglomeration characteristics, and there was no cold spot region. This confirms the view of Wang et al. [33]. This spatial pattern is consistent with the spatial structure characteristics of tourism flows, indicating that Shanghai, Nanjing, and Hangzhou are more connected with other cities. This finding supports the view of Wu et al. [47] and Zhang et al. [48].

This study contributes to the following broad literature. First, previous studies on tourist destination cooperation mainly focus on the relationship between tourist destinations. This study advances the understanding of tourist destination cooperation and holds that tourist destination cooperation should include its own competitiveness and its relationship with other tourist destinations. It has a certain complementary role to the research of tourist destination cooperation.

Second, this study constructed the evaluation model of tourist attraction cooperation and evaluated the Yangtze River Delta tourist attraction cooperation. Using a GIS analysis method, this study further analyzed the spatial characteristics of the cooperation level of tourist attractions, which can better reflect the spatial characteristics of the competitiveness of tourist attractions, cooperative relations, and the overall cooperation level. This has a significant impact on the government and enterprises to formulate the cooperation strategy of tourist attractions. Although the empirical analysis was based on a sample of tourist attractions in the Yangtze River Delta, the methodological and theoretical analysis was general and applicable around the world.

5. Conclusions

To define the connotation of cooperation of tourist attractions, this study selected key factors affecting the cooperation of tourist attractions, including two indicators reflecting the competitiveness of tourist attractions and the relationship between tourist attractions, and established the evaluation index system of tourist attraction cooperation. According to the degree of index differentiation, the entropy weight method was used to determine the index weight. In addition, we constructed the cooperation evaluation model of tourist attractions and evaluated the cooperation level of tourist attractions. By using the GIS analysis method, we further analyzed the spatial characteristics of the cooperation level of tourist attractions, which can better reflect the spatial characteristics of the competitiveness, cooperation relationship, and overall cooperation level of tourist attractions.

The following findings were obtained. First, the CETA, RITA, and CITA values of tourist attractions were generally high. The mean values revealed no significant differences between RITA, CETA, and CITA. However, by comparing the extreme values, Wuzhen had the highest RITA, which was 30.8 times higher than that of CCTV Wuxi Studios, which had the lowest RITA. The extreme multiplier of CITA and CETA was, respectively, 67.8 and 41.6 times different. Item comparison found 30 cooperative units with a difference of more than 10, accounting for 54.55% of the total. In addition, based on the constructed two-dimensional four-quadrant graph (Figure 1), Wuzhen and nine other tourist attractions located in the first quadrant had obvious advantages in competitiveness and relationship with other tourist attractions, which are distributed in Zhejiang and Jiangsu. Nanjing Zhongshan National Park, Nanxun Ancient Town, QianDao Lake, and nine other tourist attractions were located in the second quadrant and had relatively sound cooperative relations. However, compared with the tourist attractions in the first quadrant, their tourist appeal was limited. Twenty-three tourist attractions, such as Nanjing Museum, were located in the third quadrant. These tourist attractions were weak in competitiveness, attracting little attention from tourists, and had sparse relations with other tourist attractions. Finally, the fourth quadrant comprised 14 tourist attractions such as Nanxun Ancient Town. These tourist attractions had obvious competitive advantages, but their connection with other tourist attractions was slightly insufficient compared with the tourist attractions in the first quadrant. Second, the CETA, RITA, and CITA values of tourist attractions showed positive spatial autocorrelation and spatial agglomeration. CITA and RITA tended to be concentrated locally with high values, while CETA values showed overall concentration, but appeared randomly distributed locally. The RITA, CITA, and CETA values formed the high-density core group of “Shanghai Oriental Pearl Radio and TV Tower and Qing He Lane”, “Tongli Ancient Town”, and “Shanghai Oriental Pearl Radio and TV Tower, Qing He Lane, and Tongli Ancient Town, Fuzimiao (Confucius Temple) Qinhuai Scenic Area”, respectively. The high CETA values were concentrated in Shanghai, Nanjing and Hangzhou, and gradually decreased to the surrounding areas. The hotspots of CETA, RITA, and CITA values showed spatial agglomeration characteristics, and there was no cold spot region. The high values of CETA, RITA, and CITA in the northeast of Shanghai, Suzhou, Wuxi, Hangzhou, Nanjing, and Jiaxing were densely distributed. The high values of CETA and RITA decreased from the central part of the Yangtze River Delta to the south, while the high values of CITA tended to be gentle from the central part of the Yangtze River Delta to the south. The southwest tourist attractions of Zhoushan, Ningbo, Shaoxing, and Hangzhou had better tourism competitiveness, but their relationship with other tourist attractions was sparse. The spatial differentiation of an “inverted U-shape” indicated that the cooperation level, tourism competitiveness, and mutual relations of the tourist attractions were not balanced and that stable and gradual spatial transformation had not been achieved. In addition, this study is also conducive to a comprehensive understanding of the competitiveness, cooperation, and overall cooperation level of the Yangtze River Delta tourist attractions spatial characteristics. This has a significant impact on the government and enterprises to formulate the cooperation strategy of tourist attractions.

This study had some limitations. First, due to the huge number of tourist attractions, this study only included AAAA and AAAAA tourist attractions as research objects, ignoring other tourist attractions, which lacked comprehensiveness. Second, this study only considered the travel routes and tourist transaction data published on the websites of mainstream OTAs and did not consider offline transaction data. Therefore, there may be some flaws in the data. In the future, we can consider taking more tourist attractions in the Yangtze River Delta as samples to study their cooperation so as to draw more reliable conclusions. We can also use questionnaire data or direct on-site observation to analyze the spatial pattern of cooperation in tourist attractions, which can be used as the verification of GIS analysis results. It is also possible to consider using more diverse models (e.g., gravitation models) and data to study the cooperation of different types of tourism destinations so as to provide reference for improving regional tourism competitiveness.