Spatio-temporal Dataset on City Community Structure in Middle Reaches of Yangtze River (2000‒2014)

Gao, P.¹  He, D.^2*  Ning, Y. M.³

1. Shanghai Institute for Global City, Shanghai Normal University, Shanghai 200234, China;

2. School of Urban and Regional Science, East China Normal University, Shanghai 200241, China;

3. The Center for Modern Chinese City Studies, East China Normal University, Shanghai 200062, China

Abstract: On the basis of 11315 enterprise credit information system and national enterprise credit information publicity system, we obtained the data of headquarters and their branches of producer service enterprises and then constructed spatial correlation network of city agglomeration in the middle reaches of the Yangtze River (MRYR). Using community detection algorithm inclu­ding Modularity, PageRank value and alluvial diagram, we analyzed the spatio-temporal evolution characteristics of community structure of city agglomeration in the MRYR from 2000 to 2014. The dataset is archived in .xlsx format with data size of 311 KB, including process data and result data. The main results are as follows: (1) the spatial correlation network of city agglomeration in the MRYR had gradually formed an axle shape with Wuhan, Changsha and Nanchang as the radiative centers; (2) the spatial correlation network could be divided into three internal closely connected city communities, including Wuhan community, Changsha community and Nanchang community; (3) three city communities constantly adjusted their position in spatial correlation network; (4) the relationship between city communities had the characteristics of imbalance and asymmetry; (5) the administrative segmentation characteristics of city communities were obvious, and the topological structure of ??core-sub core-edge?? was formed within each city community.

Keywords: spatial correlation network; city community; spatio-temporal evolution; city agglomeration in middle reaches of the Yangtze River

DOI: https://doi.org/10.3974/geodp.2022.01.15

CSTR: https://cstr.escience.org.cn/CSTR:20146.14.2022.01.15

Dataset Availability Statement:

The dataset supporting this paper was published and is accessible through the Digital Journal of Global Change Data Repository at: https://doi.org/10.3974/geodb.2021.08.10.V1 or https://cstr.escience.org.cn/CSTR:20146.11.2021.08.10.V1.

1 Introduction

The evolution of the spatial structure of city agglomeration can be roughly divided into three development stages, including the monocentric dominance, the polycentric competition and the networked dependence and competition^[1]. Among them, the networking of city agglomeration is the highest manifestation of the dynamic flow of various resource elements in the region, and it is also an ideal urbanization model in the process of the formation and development of city agglomeration^[2,3]. City agglomeration from the perspective of network and connection is similar to the concept of ??functional area?? in economic geography^[4]. Its spatial organization emphasizes that the interaction intensity of cities within the regional boundary is closer than that outside the boundary. Furthermore, its spatial form is the aggregation of multiple cities with the central city as the radiative core in a specific region. The competition between cities has transformed into the competition between city agglomerations to a great extent^[5]. Within the city agglomeration, the relationship between alliance and competition is becoming more and more complex. Social network analysis and complex network analysis bring new concepts and analysis paradigms to geography, which provides strong support for insight into the internal spatial organization of city agglomer­ation. With the help of agglomerative subgroup algorithm such as subgroups and factions or community detection algorithm, scholars generally find that there are several city groups or city communities with ??close internal relations and sparse external relations??^[6,8]. In April 2015, the State Council approved the development plan of city agglomeration in the middle reaches of the Yangtze River, which clearly pointed out that we should build it into a new growth pole of China??s economy and promote the formation of a polycentric and networked development pattern. The city agglomeration in the middle reaches of the Yangtze River (MRYR) is a trans-provincial giant urban cluster composed of multiple urban subgroups (including Wuhan metropolitan area, Changsha-Zhuzhou-Xiangtan city agglom­eration and Poyang Lake city agglomeration). The community spatial organization pattern and its dynamic evolution trend of city agglomeration is the key to its sustainable and healthy development. Therefore, based on the headquarters and its branches of producer service enterprises, this paper constructs the dataset of spatio-temporal evolution of comm­unity structure of city agglomeration in MRYR from the perspective of urban network. The dataset can provide data support for studying and optimizing the regional development pattern.

2 Metadata of the Dataset

The metadata of the Spatio-temporal evolution dataset on community structure of city clusters in middle reaches of the Yangtze River (2000‒2014)^[9] is summarized in Table 1. It includes the dataset full name, short name, authors, year of the dataset, data format, data size, data files, data publisher, and data sharing policy, etc.

3 Methods

3.1 Data Sources

Figure 1 clarifies the database building process. First, we used the regional keyword query function of the 11315 National Enterprise Credit System and entered a few keywords including ??subsidiary??, ??branch??, and ??office?? to access branch names within study areas for the first query, while the business directory of the headquarters is obtained simultaneously. Second, we registered in the National Enterprise Credit Information Publicity System of the State Administration for Industry and Commerce to utilize the enterprise directory obtained for the second query to confirm and supplement the required information one by one. Finally,

Table 1  Metadata summary of the Spatio-temporal evolution dataset on community structure of city clusters in middle reaches of the Yangtze River (2000‒2014)

Figure 1  Building process for the producer services database

according to the Classification standard of national economy industry (GB/T 4754??2011) published by the National Bureau of Statistics, the business scope of the collected sample enterprises was classified, and thus the producer services involving six industries are segregated, including transportation, warehousing and postal services, information transmission, software and information technology services, finance, real estate, leasing and business services, and scientific research and technology services. Furthermore, we retained the samples of headquarters and branches that are in remote locations according to the locations of the enterprise headquarters and branch offices. Actually, it is in the recent two decades that producer services undergo the rapid growth and become the crucial contributor to regional network formation. Thus, we classified samples chronologically to screen out sample enterprises of 2000, 2007, and 2014 respectively. Ultimately, we obtained a total of 11,564 effective samples.

3.2 Technical Route

Firstly, taking the connected spatial units as the network nodes, the edges between the headquarters and branches of producer service enterprises were extracted, and the spatial correlation networks of urban agglomeration in the MRYR in 2000, 2007 and 2014 were constructed respectively. Secondly, the spatial correlation network of urban agglomeration in the MRYR was divided by using the Modularity index. Thirdly, this paper calculated the PageRank value of each node in each city community and drew the alluvial diagram of the dynamic evolution of city communities, and then investigated the characteristics of imbalance and asymmetry between city communities. Finally, we deeply analyzed the structural characteristics and dynamic evolution of internal city communities.

(1) Modularity: Some cities in the urban network will form several communities according to their connectivity. The nodes in one city community are relatively closer, while the links between city communities are relatively sparser. Newman et al. defined Modularity to quantitatively describe city communities in the network and measure the quality of community division^[11].

                                                     , m??n                                          (1)

where Q indicates the Modularity between 0 and 1??the closer the value to 1, the better division quality of the community structure; n is the calculated number of city communities; L is the total amount of urban links in the network; l_m is the connection quantity within the city community m; d_m is the sum number of connections associated with each node in city community m.

(2) PageRank algorithm: PageRank algorithm is an algorithm used to rank the importance of nodes in the network^[12]. The PageRank value of nodes is calculated as follows:

                                                                   (2)

where PR_i is the PageRank value; n is the number of nodes in the network; M_i is the number of nodes connected to node i; w_ij is the connection between node i and node j; D_j is the centrality of node j; d is the attenuation factor (usually 0.85).

(3) Alluvial diagram: The evolution process of community structure includes not only the changes of nodes, relationships and structures within city communities, but also the changes of relationships and positions among city communities. The alluvial diagram method proposed by Rosvall et al. can intuitively and clearly show the evolution process of city community structure^[13]. In the alluvial diagram, the name of each city community is named by the node with the largest PageRank value within the city community; The position of the city community represents its position in the network. The closer the city community is to the bottom of the alluvial diagram, the higher its position is.

4 Data Results

4.1 Data Composition

The dataset included of: (1) matrix data of spatial correlation network of city agglomeration in MRYR; (2) Modularity data of spatial correlation network of city agglomeration in MRYR; (3) PageRank value data of spatial correlation network of city agglomeration in MRYR; (4) division of city communities?? data of spatial correlation network of city agglomeration in MRYR; (5) inter-city community??s connectivity data of city agglomeration in MRYR. The dataset is archived in .xlsx format with data size of 311 KB.

4.2 Data Results

(1) The city agglomeration in the MRYR had gradually formed an axle shape with Wuhan, Changsha and Nanchang as radiative centers. With the increase of the headquarters of producer service enterprises in Wuhan, Changsha and Nanchang, the connection with surrounding cities had gradually strengthened. By investigating the first contact city of each city, it can be seen more clearly that the number of the most connected edges including the three cities has soared from 45 in 2000 to 117 in 2014. However, this reflected from the other side that the connection between many other cities except the three cities was very weak, especially the connection across the provincial administrative boundary, which will be further discussed later.

Figure 2  Spatial correlation network of city agglomeration in the MRYR

(2) The spatial correlation network of city agglomeration in the MRYR had formed three closely connected city communities. Generally speaking, if the Modularity reaches more than 0.3, it indicates that the community structure in the network is obvious. The results showed that when the division frequency of the spatial correlation network of city agglo­meration in the MRYR in 2000, 2007 and 2014 was set to 3 times, the Modularity reached the maximum in the corresponding years, which were 0.575, 0.488 and 0.45, respectively, and the effect of community division was the best, which meant that the spatial correlation network of city agglomeration in the MRYR formed three closely connected city communities. In addition, with the passage of time, the Modularity showed a decreasing trend, indicating that the internal structure of city associations was slightly loose, and the connection between city associa­tions was relatively obvious.

(3) The network status of city communities was in dynamic change. Each city community was named after the city with the highest PageRank value among it. The obtained city communities included Wuhan community, Changsha community and Nanchang community. Then, the alluvial diagram of spatio-temporal evolution of community structure of city agglomeration in the MRYR was drawn based on the Mapequation platform^[14]. The closer the city community is to the bottom of the alluvial diagram, the higher its network status will be. From 2000 to 2007, there was no change in members between city communities except that a small number of members of Changsha community became members of the other two city communities respectively; Nanchang community had the highest external connectivity and was becoming more and more important in the network. Wuhan community had surpassed Changsha community and became the second largest external connectivity community in the network. From 2007 to 2014, the change range of members among city communities increased significantly compared with the previous period, in which a considerable number of members in Nanchang community evolved into members of Wuhan community; Wuhan community once again surpassed Nanchang community and became the city community with the highest external connectivity.

(4) The relationship between city communities had the characteristics of imbalance and asymmetry. As shown in Figure 4, in 2000, the relationship between Wuhan community and Changsha community was the closest, followed by Wuhan community and Nanchang community, and the relationship between Changsha community and Nanchang community was the weakest. In 2007, Wuhan community and Nanchang community became the most closely connected pair. In 2014, the connection strength between Wuhan community and Changsha community as well as between Wuhan community and Nanchang community was similar, the interaction between city communities was active. However, the connection strength between Changsha community and Nanchang community was relatively weak. It can be seen that there was a significant imbalance in the relationship between the three city communities. Considering the leading connection flow, the relationship between city communities was not symmetrical. In the three years, Wuhan community was net outflow community, while Changsha community and Nanchang community were always net inflow communities, indicating that Wuhan community had strong external economic radiative ability and have economic externalities on the other two city communities.

(5) The administra­tive segmentation char­acteristic of city comm­unities was obvious, and the topological structure of ??core-sub core-edge?? was formed within each city community. Further research on the internal structure of urban com­munities can find that^[14], on the one hand, the provincial administrative boundary was the leading factor affecting the evolution of community structure, and the phenomenon of ??community formation?? across provinces was becoming more and more rare. On the other hand, Wuhan, Changsha and Nanchang were the core cities of city communities. Closely around the core cities were municipal nodes, which played a leading role in the secondary cores of each city community and were the important intermediary of regional spatial configuration. However, a large number of county-level nodes with remote geographical location, poor traffic conditions and weak economic foundation were distributed on the edge of each city community.

5 Discussion and Conclusion

Aiming at the phenomenon of community-based spatial organization of city agglomeration in the middle reaches of the Yangtze River (MRYR), this study compiles the spatio-temporal evolution of community structure dataset of city agglomeration in the MRYR (2000‒2014). This dataset not only helps to form a new understanding of the spatial structure and organization of city agglomerations from the perspective of urban network, but also provides basic data and reference for the formulation of regional coordinated development policies. The data results showed that during the study period, the spatial correlation network of urban agglomeration in the MRYR had become increasingly close, forming an axle shape with Wuhan, Changsha and Nanchang as the radiating centers. In each year, the spatial correlation network of urban agglomerations in the MRYR was divided into three city communities: Wuhan community, Changsha community and Nanchang community. The status of three city communities in the network constantly adjusted, and the relationship between different city communities was unbalanced and asymmetric. In addition, with the passage of time, the phenomenon of ??community formation?? across provinces disappeared, city communities were obviously divided by administrative boundaries, and a ring hierarchical structure was formed within each one.

This study digs the directory of enterprise headquarters and their branch enterprises from network big data, uses the national enterprise credit information publicity system to verify the accuracy of the data and further supplement effective information. It is a beneficial attempt to combine enterprise network big data with official data, which not only reflects the timeliness of the data, but also ensures the reliability of the data. In addition, this study also combines social network analysis with spatial analysis, reveals the phenomenon of community spatial organization of city agglomeration from the perspective of urban network, and can provide a more scientific empirical analysis path for relevant research. It should be noted that the enterprise headquarters branch data in this dataset is the number of invested enterprises. In the follow-up work, it is necessary to mine the investment quota data to build a spatial association network more accurately.

Author Contributions

He, D. and Ning, Y. M. set up the framework for the development of the dataset; Gao, P. and He, D. completed the data collection, processing and verification; Gao, P. completed the data operation and wrote the paper.

Conflicts of Interest

The authors declare no conflicts of interest.

References

[1]      Scott, A. J. Global City-Regions: Trends, Theory, Policy [M]. New York: Oxford University Press, 2001.

[2]      Nian, F. H., Yao, S. M., Chen, Z. G. The preliminary study on the network organization in urban agglomeration [J]. Scientia Geographica Sinica, 2002, 22(5): 568?C573.

[3]      Pan, F. H., Fang, C., Li, X. D. The progress and prospect of research on Chinese city network [J]. Scientia Geographica Sinica, 2019, 39(7): 1093?C1101.

[4]      Anderson, W. Economic Geography [M]. New York: Routledge, 2012.

[5]      Zhang, F., Ning, Y. M., Lou, X. Y. Competitiveness and regional inequality of China??s mega-city regions [J]. Geographical Research, 2019, 38(7): 1664?C1677.

[6]      Fang, D. C., Sun, M. Y. The reconstruction of the spatial structure of the Yangtze River Delta city group in the high-speed rail era: based on the social network analysis [J]. Economic Geography, 2015, 35(10): 50?C56.

[7]      Zhang, W. Y., Derudder, B., Wang, J. E., et al. Regionalization in the Yangtze River Delta, China, from the perspective of inter-city daily mobility [J]. Regional Studies, 2018, 52(4): 528?C541.

[8]      Gao, P., He, D., Sun, Z. J., et al. Characterizing functionally integrated regions in Central Yangtze River Megaregion from a city-network perspective [J]. Growth and Change, 2020, 51: 1357?C1379.

[9]      Gao, P., He, D., Ning, Y. M. Spatio-temporal evolution of community structure dataset of city agglomeration in the middle reaches of the Yangtze River (2000‒2014) [J/DB/OL]. Digital Journal of Global Change Data Repository, 2021. https://doi.org/10.3974/geodb.2021.08.10.V1. https://cstr.escience.

org.cn/CSTR:20146.11.2021.08.10.V1.

[10]    GC dataPR Editorial Office. GC dataPR data sharing policy [OL]. https://doi.org/10.3974/dp.policy. 2014.05 (Updated 2017).

[11]    Newman, M. E. Finding community structure in networks using the eigenvectors of matrices [J]. Physical Review E: Statistical Nonlinear & Soft Matter Physics, 2006, 74(3): 1?C22.

[12]    Brin, S., Page, L. Reprint of: the anatomy of a large-scale hypertextual web search engine [J]. Computer Networks, 2012, 56(18): 3825?C3833.

[13]    Rosvall, M., Bergstrom, C. T. Mapping change in large networks [J]. PLoS One, 2010, 5(1): 1?C7.

[14]    Gao, P., He, D., Ning, Y. M., et al. Community structure and proximity mechanism of city clusters in middle reaches of the Yangtze River: based on producer service firms?? network [J]. Scientia Geographica Sinica, 2019, 39(4): 578?C586.