Skip to Main Content
Publication Cover
Journal

Tourism Geographies

An International Journal of Tourism Space, Place and Environment

Latest Articles

ABSTRACT

The research shift from central place theory to central flow theory has demonstrated the influence of information technology on cities. The study explores this shift at the city scale of urban catering industry in Nanjing, China. A comprehensive set of indicators of E-WoM for the catering industry has been established. Based on this, the spatial distribution patterns of catering industry in Nanjing, China have been discussed to examine the relationship of restaurants distribution and the central place theory or the central flow theory using a data analytical approach. The results revealed that the spatial distribution of restaurants’ density in Nanjing follows a traditional central place theory, but the spatial distribution of restaurants’ E-WoM in Nanjing rather follows the central flow theory. In addition, different characteristics could be found in different types of restaurants’ E-WoM distribution. Mainstream cuisine follows central place theory but is inconsistent in urban business districts, which demonstrate some characteristics of central flow theory. The distribution of E-WoM of non-mainstream cuisine is similar to the E-WoM distribution of overall restaurants, showing a central flow pattern. Finally, the implications of the study are drawn.

摘要

从中心地理论到中心流理论的研究转向表明了信息技术对城市的影响。本研究探讨了中国南京城市餐饮业的这种转变。研究建立了一套基于网络口碑度的餐饮评价指标。基于此, 对南京餐饮业的空间分布格局进行了研究, 探讨了餐饮分布与中心地理论或中心流理论的关系。结果表明, 南京餐饮的空间分布遵循传统的中心地理论, 但餐饮的网络口碑度遵循中心流理论。此外, 不同类型的餐饮呈现不同的网络口碑分布。主流餐饮主要遵循中心地理论, 但在城市商业区却部分体现了中心流理论的特点。非主流餐饮网络口碑的分布则类似于餐饮业整体的网络口碑分布, 呈现出中央流模式。本文最后提出一些实践对策。

1. Introduction

Central place theory indicates that residents tend to visit the nearest city to buy goods and services when they are restricted by market, transportation, and administration (Christaller, 1966 Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. [Google Scholar]). Urban development presents a strict hierarchal distribution. However, the development of regions and cities has no longer been analyzed in a closed system because of globalization. Cities became closely linked due to the development of a comprehensive transportation network, therefore leading to the formation of a complicated urban network. In this process, the development and extensive use of information communication technology (ICT) have accelerated the exchanges of information, population, capital, goods, and other elements among cities (Zhen, Yu, Wang, & Zhao, 2012 Zhen, F., Yu, Y., Wang, X., & Zhao, L. (2012). The spatial agglomeration characteristics of automotive service industry: A case study of Nanjing. Scientia Geographica Sinica, 32(10), 12001208. [In Chinese]. [Google Scholar]). Castells (1989 Castells, M. (1989). The informational city: Information technology, economic restructuring, and the urban-regional process. Oxford: Blackwell. [Google Scholar]) focused on the emergence of a new technology paradigm with the influence of ICT and proposed the concept of ‘flow space.’ He also distinguished flow space and place space, as well as expanded the flow space from pure virtual technology space to geographical space and social network scale (Castells, 1996 Castells, M. (1996). Rise of the network society: The information age: Economy, society and culture. Massachusetts: Blackwell Publishers. [Google Scholar]). Taylor (2001 Taylor, P. J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181194.[Crossref], [Web of Science ®] [Google Scholar]), Taylor, Evans, and Pain (2008 Taylor, P. J., Evans, D. M., & Pain, K. (2008). Application of the interlocking network model to mega-city-regions: Measuring polycentricity within and beyond city-regions. Regional Studies, 42(8), 10791093.[Taylor & Francis Online], [Web of Science ®] [Google Scholar]), and Taylor, Hoyler, and Verbruggen (2010 Taylor, P. J., Hoyler, M., & Verbruggen, R. (2010). External urban relational process: Introducing central flow theory to complement central place theory. Urban Studies, 47(13), 28032818.[Crossref], [Web of Science ®] [Google Scholar]) further found that the level of spatial distribution of ‘Central Place’ theory has evidently been changed by ‘Central Flow’ theory. They suggested that the vibrant city is the center of the spatial flow, and the complexity of the flow is evident in non-local characteristics of the city population, goods and ideas. The hinterland of the city may not be adjacent to a certain range of space, but the distribution tends to be fragmented.

In order to adapt to this trend of fast change, scholars have gradually found the new breakthrough in paradigm, content and method of research, which outstandingly uses network information data to study flow space and place space (Zhen, Wang, & Wei, 2015 Zhen, F., Wang, B., & Wei, Z. (2015). The rise of the internet city in china: Production and consumption of internet information. Urban Studies, 52(13), 23132329.[Crossref], [Web of Science ®] [Google Scholar]). They focused on the regional urban network and their flow patterns by analyzing the Network infrastructure data (e.g. backbone network bandwidth and network number, Internet domain name or IP address) ( Zook, 2001 Zook, M. A. (2001). Old hierarchies or new networks of centrality: The global geography of the Internet content market. American Behavioral Scientist, 44(10), 16791696.[Crossref], [Web of Science ®] [Google Scholar]; Zook & Graham, 2007 Zook, M. A., & Graham, M. (2007). Mapping digiplace: Geocoded internet data and the representation of place. Environment and Planning B: Planning and Design, 34(3), 466482.[Crossref], [Web of Science ®] [Google Scholar]), social network data (e.g. Twitter, Weibo) (Naaman, Zhang, & Brody, 2012 Naaman, M., Zhang, A. X., & Brody, S. (2012). On the study of diurnal urban routines on Twitter. Sixth international AAAI conference on Weblogs and Social Media, Dublin. [Google Scholar]; Zhen et al., 2012 Zhen, F., Yu, Y., Wang, X., & Zhao, L. (2012). The spatial agglomeration characteristics of automotive service industry: A case study of Nanjing. Scientia Geographica Sinica, 32(10), 12001208. [In Chinese]. [Google Scholar]), and mobile phone data (Kang, Zhang, & Ma, 2012 Kang, C., Zhang, Y., Ma, X., & Liu, Y. (2012). Inferring properties and revealing geographical impacts of intercity mobile communication network of China using a subnet data set. International Journal of Geographical Information Science, 27 (3) 118.[Web of Science ®] [Google Scholar]; Krings, Calabrese, & Ratti, 2009 Krings, G., Calabrese, F., & Ratti, C., & Blondel, V. D. (2009). Urban gravity: A model for inter-city telecommunication flows. Journal of Statistical Mechanics: Theory and Experiment, (7), 18. [Google Scholar]). However, these theories and methods were mainly used to explain interactions between the cities, and research on the spatial distribution of production factors within the city is limited. Several questions remain unexplored. Do the spatial patterns of production factors within a city follow the rules of central place theory? Can the central flow theory be used to explain new phenomenon? Does the new phenomenon demonstrate the features of central flow theory at regional scale? Can traditional data on urban interior space support and explain the new phenomenon? Is there any new dataset to explain this phenomenon? Therefore, these research questions need further exploration.

Due to the fast development of information technology, especially the ICT platforms such as dianping.com (a Chinese website that allows users to buy coupons and post review comments on food and dining experiences), consumers’ behavior of dinning in restaurants based on the central place theory has changed (Bei, Chen, Rha, & Widdows, 2004 Bei, Lien-Ti, Chen, Etta Y. I., Rha, Jong-Youn, & Widdows, Richard (2004). Consumers’ online information search for a new restaurant for dining-out. Journal of Foodservice Business Research, 6(3), 1536.[Taylor & Francis Online] [Google Scholar]). On the one hand, the information about restaurants on the platform has given consumers more choices, providing more dining flexibility and expanded the scope of their activities (Qin, Zhen, Zhu, & Guang-Liang, 2014 Qin, X., Zhen, F., Zhu, S. J., & Guang-Liang, X. I. (2014). Spatial pattern of catering industry in Nanjing urban area based on the degree of public praise from internet: A case study of dianping.com. Scientia Geographica Sinica, 34(7), 810817. [In Chinese]. [Google Scholar]). On the other hand, comments made by other consumers on the review websites (positive or negative) also change the spatial and temporal behavior of consumers’ dining activities. It is precisely these forces (Eckardt, 2008 Eckardt, F. (2008). Media and urban space. In F. Eckardt (Eds.), Media and urban space: Understanding, investigating and approaching mediacity (pp. 79). Berlin: Frank & Timmer Gmbh. [Google Scholar]) that may lead to the change of distribution of those restaurants that traditionally followed urban business districts, creating more mobility for consumers. As can be seen, after the emergence of new Internet platforms, the spatial distribution of the traditional catering industry might change. Traditional central place theory that reflects the rule of spatial distribution of urban catering industry might change the decision-making and consuming behavior, and the spatial distance between organizations might increase. Therefore, this paper uses urban catering space which has significantly been influenced by new technology as a study object, and explores Nanjing, China as the case area. The paper tries to explore the application of central place theory and central flow theory in production elements within the city with the influence of new technology. Current studies on catering industry mainly adopt traditional methods, which include the use of spatial and statistical data such as catering location, scale, price, spatial and statistical data and explore the spatial distribution pattern and its determinants at urban macro level; or the use of consumer word-of-mouth information through questionnaires to evaluate the service quality of catering providers with a relatively limited sample. However, these types of data, which reflect the spatial information, lack information on catering flow space, which is actually reflected by consumers’ and catering providers’ behavior with new platforms. In recent years, E-WoM which refers to electronic word of mouth (King, Racherla, & Bush, 2014 King, R. A., Racherla, P., & Bush, V. D. (2014). What we know and don't know about online word-of-mouth: A review and synthesis of the literature. Journal of Interactive Marketing, 28(3), 167183.[Crossref], [Web of Science ®] [Google Scholar]) has gradually become an important factor influencing consumer behavior (Berger, 2014 Berger, J. (2014). Word of mouth and interpersonal communication: A review and directions for future research. Journal of Consumer Psychology, 24(4), 586607.[Crossref], [Web of Science ®] [Google Scholar]), and can be used to reflect urban catering providers’ attractiveness to consumer behavior without the constraint of spatial distance.

This paper uses Nanjing, China as a case study, using E-WoM from online reviews, to discuss the spatial pattern of urban catering industry and to explore the application and new features of central place theory and central flow theory at urban scale. To be specific, the research questions include: (1) what is the E-WoM of urban catering industry in Nanjing? (2) What is the spatial pattern of urban catering providers and its relationship with central flow theory in Nanjing? (3) What is the spatial pattern of E-WoM of urban providers and its relationship with central place theory, and whether it fits with the features of regional urban network of central flow theory in Nanjing? (4) Do different types of catering providers’ E-WoM reflect the providers’ distribution pattern?

2. Literature review

2.1. From central place theory to central flow theory

Since Christaller's central place theory was proposed (Christaller, 1966 Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. [Google Scholar]), location and spatial distribution pattern of urban service industry have always been a research focus. The research scope mainly involves Commercial Service Industry and Productive Service Industry. Christaller (1966 Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. [Google Scholar]) suggests that customers will travel only to the nearest central place that provides goods and services that they demand, and goods are purchased from the closest place. Classical location theory emphasizes on economic factors and suggest that labor costs and market opportunities are critical. Ohlin (1993 Ohlin, B. (1993). 1933 and 1977–some expansion policy problems in cases of unbalanced domestic and international economic relations. American Economic Review, 83(6), 1017. [Google Scholar]) suggests that the Neo-classical trade theory introduces exogenous factors such as natural resources, labor and technology to the location selection. New economic geography theory suggests that connections to other businesses and traffic costs lead to a cluster of businesses (Krugman, 1991 Krugman, P. R. (1991). Increasing returns and economic geography.Southern Economic Journal, 99(3), 483500. [Google Scholar]).

Globalization has made the cities no longer a closed local system separated from others, as a comprehensive transportation network has made cities and facilities closely linked, forming a dynamic regional urban network (Zhen, Liu, & Liu, 2007 Zhen, F., Liu, X. X., & Liu, H. (2007). Regional urban network influenced by information technology: New directions of urban studies. Human Geography, 22(2), 7680. [In Chinese]. [Google Scholar]). Researchers have now modified the ideal hierarchical central place model. Through a study on Northeast America, Gottmann (1961 Gottmann, J. (1961). Megalopolis, the urbanized northeastern seaboard of the United States. New York, NY: The Twentieth Century Fund. [Google Scholar]) finds ‘horizontal and non-hierarchical, polycentric’ network pattern in urban areas, which contains two or more than two separated cities, and is connected with fast and convenient traffic corridors, emphasizing the node role of cities in the network rather than the central role (Batten, 1995 Batten, D. (1995). Network cities: Creative urban agglomerations for the 21st century. Urban Studies, 32(2), 313327.[Crossref], [Web of Science ®] [Google Scholar]). Camagni and Salone (1993 Camagni, R., & Salone, C. (1993). Network urban structures in northern Italy: Elements for a theoretical framework. Urban Studies, 30(30), 10531064.[Crossref] [Google Scholar]) suggest that different from critical factors in central place theory, such as economic scale, production scale, demand and market scale; the micro economy and local enterprises network is the driving factor of urban regions (Batten, 1995 Batten, D. (1995). Network cities: Creative urban agglomerations for the 21st century. Urban Studies, 32(2), 313327.[Crossref], [Web of Science ®] [Google Scholar]; Camagni & Salone, 1993 Camagni, R., & Salone, C. (1993). Network urban structures in northern Italy: Elements for a theoretical framework. Urban Studies, 30(30), 10531064.[Crossref] [Google Scholar]; Knaap, 2002 Knaap, G. A. van der. (2002). Stedelijke bewegingsruimte, over veranderingen in stad en land. The Hague: Sdu Uitgevers. [Google Scholar]; Meijers, 2005 Meijers, E. (2005). Polycentric urban regions and the quest for synergy: Is a network of cities more than the sum of the parts? Urban Studies, 42(42), 765781.[Crossref] [Google Scholar]).

In recent decades, the rapid development and wide application of ICT has brought the speedy exchange of factors at spatial and temporal levels, such as information, population, capital and goods in the city (Schwanen, Dijst, & Kwan, 2006 Schwanen, T., Dijst, M., & Kwan, M. P. (2006). Introduction-the internet, changing mobilities, and urban dynamics. Urban Geography, 27(7), 585589.[Taylor & Francis Online], [Web of Science ®] [Google Scholar]), which has continuously expanded the scope of urban production and residents’ activities, and, as a result, the urban network model is continuously improved. Castells (1989 Castells, M. (1989). The informational city: Information technology, economic restructuring, and the urban-regional process. Oxford: Blackwell. [Google Scholar]) was one of the first who focused on the emergence of new technology paradigm under the influence of information technology, emphasizing the importance of information processing and process innovation, and put forward the concept of ‘flow space’ based on information technology. Castells (1996 Castells, M. (1996). Rise of the network society: The information age: Economy, society and culture. Massachusetts: Blackwell Publishers. [Google Scholar]) further distinguishes flow space and place space, and expands the flow space from pure virtual technology space to geographical space and social network scale. Therefore, based on theories of Camagni and Salone (1993 Camagni, R., & Salone, C. (1993). Network urban structures in northern Italy: Elements for a theoretical framework. Urban Studies, 30(30), 10531064.[Crossref] [Google Scholar]) and Castells (1996 Castells, M. (1996). Rise of the network society: The information age: Economy, society and culture. Massachusetts: Blackwell Publishers. [Google Scholar]), Taylor (2001 Taylor, P. J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181194.[Crossref], [Web of Science ®] [Google Scholar]) and Taylor et al. (2008 Taylor, P. J., Evans, D. M., & Pain, K. (2008). Application of the interlocking network model to mega-city-regions: Measuring polycentricity within and beyond city-regions. Regional Studies, 42(8), 10791093.[Taylor & Francis Online], [Web of Science ®] [Google Scholar], 2010 Taylor, P. J., Hoyler, M., & Verbruggen, R. (2010). External urban relational process: Introducing central flow theory to complement central place theory. Urban Studies, 47(13), 28032818.[Crossref], [Web of Science ®] [Google Scholar]) further use the interlocking network model to evaluate the relationship between cities. They find that the level of spatial distribution of ‘central place’ theory has obviously been changed by the ‘central flow’ theory, such that the vibrant city is the center of the spatial flow, and stress the complexity of the flow, which is non-local characteristics of the city population, goods and ideas. The hinterland of the city may not be adjacent to a certain range of space, but the distribution tends to be fragmented.

However, urban network model or central flow theory cannot replace central place theory, as the former represents the spatial characteristics of urban regions in the service economy while the latter explains the relationship of cities in industrial economy (Batten, 1995 Batten, D. (1995). Network cities: Creative urban agglomerations for the 21st century. Urban Studies, 32(2), 313327.[Crossref], [Web of Science ®] [Google Scholar]; Camagni & Salone, 1993 Camagni, R., & Salone, C. (1993). Network urban structures in northern Italy: Elements for a theoretical framework. Urban Studies, 30(30), 10531064.[Crossref] [Google Scholar]; Knaap, 2002 Knaap, G. A. van der. (2002). Stedelijke bewegingsruimte, over veranderingen in stad en land. The Hague: Sdu Uitgevers. [Google Scholar]; Meijers, 2005 Meijers, E. (2005). Polycentric urban regions and the quest for synergy: Is a network of cities more than the sum of the parts? Urban Studies, 42(42), 765781.[Crossref] [Google Scholar]). To be specific, with the background of globalization and informatization, the following features appear in regional space: (1) The role of the central area has not disappeared, and the urban or regional space still has a certain hierarchy; (2) cities with large population and commodities might not have a high grade in the region, while cities that act as the hub or important node function of regional population and commodity flow network will become a higher grade center; (3) the spatial gathering effect of the traditional geographical factors is weakened, the dispersion effect is enhanced, the regional center develops independently, and the fragmentation of the urban hinterland becomes obvious.

The urban network model and central flow theory originate from central place theory, and both are expansions and improvements of central place theory at different stages of urbanization and technology development, and explore the relationship of urban production factors and spatial distance in a certain region (Smith, 1985 Smith, S. L. J. (1985). Location patterns of urban restaurants. Annals of Tourism Research, 12(4), 581602.[Crossref], [Web of Science ®] [Google Scholar]). However, these theories are mainly used to explain interactions between the cities, and research on the spatial distribution of production factors within a city is limited or, in other words, there is need to study further these features of central flow theory which can be found within cities.

2.2. The spatial distribution of the catering industry

First, a definition of the catering industry is provided before any further discussion. Catering refers to ‘the provision of food and beverages away from home’ (Davis, Lockwood, & Stone, 1998 Davis, B., Lockwood, A., & Stone, S. (1998). Food and beverage management (3rd ed.). Oxford: Butterworth-Heinemann. [Google Scholar]). Fusi, Guidetti, and Azapagic (2016 Fusi, A., Guidetti, R., & Azapagic, A. (2016). Evaluation of environmental impacts in the catering sector: The case of pasta. Journal of Cleaner Production, 132(20), 146160.[Crossref] [Google Scholar]) divide catering sectors into profit and non-profit activities. Bourlakis and Weightman (2004 Bourlakis, M. A., & Weightman, P. W. H. (2004). Food supply chain management, Oxford: Blackwell Publishing. [Google Scholar]) suggest that the profit catering sectors include restaurants, fast-food chain outlets, cafes, takeaways, pubs, leisure and travel catering outlets, while the non-profit sectors refer to catering outlets for business, education and health care (Fusi et al., 2016 Fusi, A., Guidetti, R., & Azapagic, A. (2016). Evaluation of environmental impacts in the catering sector: The case of pasta. Journal of Cleaner Production, 132(20), 146160.[Crossref] [Google Scholar]). In this research, only the profit sectors have been considered, therefore the study objects in this research website include restaurants, cafes, fast-food chain outlets, cafes and takeaways which are available from the online platform of dianping.com.

With the guidance of traditional central place theory, the spatial distribution of food and beverage industry is a research topic that scholars pay close attention to. Similar to the spatial distribution of other service industries (Zhen et al., 2012 Zhen, F., Yu, Y., Wang, X., & Zhao, L. (2012). The spatial agglomeration characteristics of automotive service industry: A case study of Nanjing. Scientia Geographica Sinica, 32(10), 12001208. [In Chinese]. [Google Scholar]; Zhou, Zhen, Yu, & Jiang, 2010 Zhou, K. H., Zhen, F., Yu, Y., & Jiang, Y. H. (2010). A research on the processes and patterns of spatial agglomeration of financial services in urban central area: A case study of Kuiwen district, Weifang city. Human Geography, 25(6), 6267. [In Chinese]. [Google Scholar]), most scholars believe that the distribution of urban restaurants follows the rule of falling from the city's Central Business Districts (CBD) to the periphery areas. There are aggregated clusters and territorial distribution within each area (Gwohshiung, Teng, Chen, & Opricovic, 2002 Gwohshiung, T., Teng, M. H., Chen, J. J., & Opricovic, S. (2002). Multicriteria selection for a restaurant location in Taipei. International Journal of Hospitality Management, 21(2), 171187.[Crossref] [Google Scholar]; Zhang & Xu, 2009 Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. [Google Scholar]), and generally speaking, these are closely related to traffic routes (Hu & Zhang, 2002 Hu, Z. Y., & Zhang, Z. G. (2002). An analysis about the spatial distribution of hotels in urban area: Take Nanjing city as a case. Economic Geography, 22(1), 106110. [In Chinese] [Google Scholar]). Meanwhile, the level of restaurants and the development level of business districts show high coupling (Zhang & Xu, 2009 Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. [Google Scholar]). Muller and Inman (1994 Muller, C. C., & Inman, C. (1994). The geodemographics of restaurant development. Cornell Hospitality Quarterly, 35(3), 8895.[Crossref] [Google Scholar]) believe that urban restaurants mainly scatter in the urban retail areas, and are customer-oriented. Restaurants that are located near each other form a broad central system, and are linked with urban retail and residential areas. Many factors have been found critical to the spatial distribution of restaurants at the city level, such as traffic accessibility (Melaniphy, 1992 Melaniphy, J. C. (1992). Restaurant and fast-food site selection. New York, NY: Wiley. [Google Scholar]; Austin et al., 2005 Austin, S. B., Melly, S. J., Sanchez, B. N., Patel, A., Buka, S., & Gortmaker, S. L. (2005). Clustering of fast-food restaurants around schools: A novel application of spatial statistics to the study of food environments. American Journal of Public Health, 95(9), 15751581.[Crossref], [PubMed], [Web of Science ®] [Google Scholar]), urban spatial pattern, economic development (Liang, 2007 Liang, L. (2007). The distribution in space of urban catering and its factors: Xi'an as an example. Journal of Northwest University (Natural Science Edition), 37(6), 925930. [In Chinese]. [Google Scholar]; Shu, Wang, Sun, Liu, & Xiao, 2012 Shu, S., Wang, R., Sun, Y., Liu, J., & Xiao, L. (2012). Spatial distribution of urban catering industry and its influenced factors: A case study of Xiamen City. Tropical Geography, 32(2), 134140. [Google Scholar]; Zhang & Xu, 2009 Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. [Google Scholar]), population density (Liang, 2007 Liang, L. (2007). The distribution in space of urban catering and its factors: Xi'an as an example. Journal of Northwest University (Natural Science Edition), 37(6), 925930. [In Chinese]. [Google Scholar]; Shu et al., 2012 Shu, S., Wang, R., Sun, Y., Liu, J., & Xiao, L. (2012). Spatial distribution of urban catering industry and its influenced factors: A case study of Xiamen City. Tropical Geography, 32(2), 134140. [Google Scholar]), infrastructure (Austin et al., 2005 Austin, S. B., Melly, S. J., Sanchez, B. N., Patel, A., Buka, S., & Gortmaker, S. L. (2005). Clustering of fast-food restaurants around schools: A novel application of spatial statistics to the study of food environments. American Journal of Public Health, 95(9), 15751581.[Crossref], [PubMed], [Web of Science ®] [Google Scholar]; Hu & Zhang, 2002 Hu, Z. Y., & Zhang, Z. G. (2002). An analysis about the spatial distribution of hotels in urban area: Take Nanjing city as a case. Economic Geography, 22(1), 106110. [In Chinese] [Google Scholar]), market demand; Hu & Zhang, 2002 Hu, Z. Y., & Zhang, Z. G. (2002). An analysis about the spatial distribution of hotels in urban area: Take Nanjing city as a case. Economic Geography, 22(1), 106110. [In Chinese] [Google Scholar]), cultural factors (Liang, 2007 Liang, L. (2007). The distribution in space of urban catering and its factors: Xi'an as an example. Journal of Northwest University (Natural Science Edition), 37(6), 925930. [In Chinese]. [Google Scholar]; Zhang & Xu, 2009 Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. [Google Scholar]), land availability (Shu et al., 2012 Shu, S., Wang, R., Sun, Y., Liu, J., & Xiao, L. (2012). Spatial distribution of urban catering industry and its influenced factors: A case study of Xiamen City. Tropical Geography, 32(2), 134140. [Google Scholar]), density of competitors (Litz & Rajaguru, 2008 Litz, R. A., & Rajaguru, G. (2008). Does small store location matter? A test of three classic theories of retail location. Journal of Small Business & Entrepreneurship, 21(4), 477492.[Taylor & Francis Online] [Google Scholar]), number of complementary stores (Schaefer, Luke, & Green, 1996 Schaefer, A. D., Luke, R. H., & Green, J. (1996). Attitudes of restaurant site selection executives toward various people magnets. Journal of Restaurant & Foodservice Marketing, 1(3), 114.[Taylor & Francis Online] [Google Scholar]), land price (Hu & Zhang, 2002 Hu, Z. Y., & Zhang, Z. G. (2002). An analysis about the spatial distribution of hotels in urban area: Take Nanjing city as a case. Economic Geography, 22(1), 106110. [In Chinese] [Google Scholar]) and urban tourism activities (Liang, 2007 Liang, L. (2007). The distribution in space of urban catering and its factors: Xi'an as an example. Journal of Northwest University (Natural Science Edition), 37(6), 925930. [In Chinese]. [Google Scholar]). Other factors at the restaurant level include scale and cost (Smith, 1995; Shu et al., 2012 Shu, S., Wang, R., Sun, Y., Liu, J., & Xiao, L. (2012). Spatial distribution of urban catering industry and its influenced factors: A case study of Xiamen City. Tropical Geography, 32(2), 134140. [Google Scholar]; Timor & Sipahi, 2005 Timor, M., & Sipahi, S. (2005). Fast-food restaurant site selection factor evaluation by the Analytical Hierarchy Process. The Business Review Cambridge, 4(1), 161167. [Google Scholar]), type of restaurant (Teller & Reutterer, 2008 Teller, C., & Reutterer, T. (2008). The evolving concept of retail attractiveness: What makes retail agglomerations attractive when customers shop at them? Journal of Retailing & Consumer Services, 15(3), 127143.[Crossref] [Google Scholar]; Zhang & Xu, 2009 Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. [Google Scholar]), parking facilities) and identity (Melaniphy, 1992 Melaniphy, J. C. (1992). Restaurant and fast-food site selection. New York, NY: Wiley. [Google Scholar]; Schaefer et al., 1996 Schaefer, A. D., Luke, R. H., & Green, J. (1996). Attitudes of restaurant site selection executives toward various people magnets. Journal of Restaurant & Foodservice Marketing, 1(3), 114.[Taylor & Francis Online] [Google Scholar]).

It can be seen that current research on restaurants’ spatial distribution pattern mainly focus on scales and levels of restaurants, and the distribution patterns of ‘circular diminishing, closely related to traffic arteries, coupling with retail trades’ can be explained by central place theory. However, there is very limited research on restaurants’ distribution and Central Flow Theories considering the influence of ICT. Researchers mainly discuss the factors that influence restaurants’ location at the city level, but hardly consider the restaurants quality. Indeed, except the factors such as scale, cost, type, parking facilities and identity, other factors such as customer preference, type of food, atmosphere, service, decoration, reputation, brand and value may all influence the choice of restaurants (Baek, Ham, & Yang, 2006 Baek, S. H., Ham, S., & Yang, I. S. (2006). A cross-cultural comparison of fast food restaurant selection criteria between Korean and Filipino college students. International Journal of Hospitality Management, 25(4), 683698.[Crossref] [Google Scholar]; Kincaid, Baloglu, Mao, & Busser, 2010 Kincaid, C., Baloglu, S., Mao, Z., & Busser, J. (2010). What really brings them back?: The impact of tangible quality on affect and intention for casual dining restaurant patrons. International Journal of Contemporary Hospitality Management, 22(2), 209220(12).[Crossref], [Web of Science ®] [Google Scholar]; Kivela, 1997 Kivela, J. J. (1997). Restaurant marketing: Selection and segmentation in Hong Kong. International Journal of Contemporary Hospitality Management, 9(3), 116123.[Crossref] [Google Scholar]; Yüksel & Yüksel, 2003 Yüksel, A., & Yüksel, F. (2003). Measurement of tourist satisfaction with restaurant services: A segment-based approach. Journal of Vacation Marketing, 9(1), 5268.[Crossref] [Google Scholar]). These are critical factors influencing the spatial distribution of urban restaurants.

2.3. E-WoM

Hospitality is part of the service industry. Researchers suggest consumers often rely heavily on friends and family when they try a new type of service, and therefore, word of mouth (WoM) as an influential factor can influence consumer decision-making (Anderson, 1998 Anderson, E. W. (1998). Customer satisfaction and word of mouth. Journal of Service Research, 1(1), 517.[Crossref] [Google Scholar]; Brown, Broderick, & Lee, 2007 Brown, J., Broderick, A. J., & Lee, N. (2007). Word of mouth communication within online communities: Conceptualizing the online social network. Journal of Interactive Marketing, 21(3), 220.[Crossref], [Web of Science ®] [Google Scholar]).

Traditionally, marketing communication is passed and benefited through family and friends (Brown & Reingen, 1987 Brown, J. J., & Reingen, P. H. (1987). Social ties and word-of-mouth referral behavior. Journal of Consumer Research, 14(3), 350362.[Crossref], [Web of Science ®] [Google Scholar]). Nowadays, the Internet is regarded as the most innovative technology over the last few decades (Beldad, Jong, & Steehouder, 2010 Beldad, A., Jong, M. D., & Steehouder, M. (2010). How shall I trust the faceless and the intangible? A literature review on the antecedents of online trust. Computers in Human Behavior, 26(5), 857869.[Crossref], [Web of Science ®] [Google Scholar]). With the fast development of Internet, particularly the web 2.0 which allows customers to share their experiences and comments online, WoM has now shifted to E-WoM, expanding networking from family and friends to people who are connected online. Hart and Blackshaw (2006 Hart, C., & Blackshaw, P. (2006). Internet INFERNO. Marketing Management, 15(1), 1825. Available from http://www.ehis.ebscohost.com [Accessed 26 January 2012]. [Google Scholar]) suggest that compared with traditional word of mouth, ‘Word of Web’ can include a social network that spans globally. E-WoM provides browsers with wider information, their views on the tangible products based on their memories of past time leisure experiences in tourism cities, hotels, airlines and restaurants (Yoon, 2009 Yoon, S. (2009). The effects of electronic word-of-mouth systems (EWOMS) on the acceptance of recommendation (Dissertation Abstracts International Section A, 69, 7-A). PsycINFO, EBSCOhost. [Google Scholar]). Senecal and Nantel (2004 Senecal, S., & Nantel, J. (2004). The influence of online product recommendations on consumers online choices. Journal of Retailing, 80, 159169.[Crossref], [Web of Science ®] [Google Scholar]) suggest that consumers show a tendency of making purchases following online recommendations. Zhu and Zhang (2012 Zhu, F., & Zhang, X. (2012). Impact of online consumer reviews on sales: The moderating role of product and consumer characteristics. Journal of Marketing, 74(2), 133148.[Crossref], [Web of Science ®] [Google Scholar]) also suggest that online recommendations significantly influence sales. Gretzel and Yoo's (2008 Gretzel, U., & Yoo, K. H. (2008). Use and impact of online travel reviews. Information and communication technologies in tourism, Enter 2008. Proceedings of the international conference in Innsbruck, (Vol. 26, pp. 3546). Vienna: DBLP.[Crossref] [Google Scholar]) research suggest that three-quarters of travelers have taken consideration of online consumer reviews when planning their holiday journeys.

E-WoM has become popular and become a key source of information about specific products (Hollenstein & Purves, 2010 Hollenstein, L., & Purves, R. (2010). Exploring place through user-generated content: Using flickr to describe city cores. Journal of Spatial Information Science, 1(1), 2148. [Google Scholar]; Litz & Rajaguru, 2008 Litz, R. A., & Rajaguru, G. (2008). Does small store location matter? A test of three classic theories of retail location. Journal of Small Business & Entrepreneurship, 21(4), 477492.[Taylor & Francis Online] [Google Scholar]). Sparks and Browning (2010 Sparks, B. A., & Browning, V. (2010). Complaining in cyberspace: The motives and forms of hotel guests’ complaints online. Journal of Hospitality Marketing & Management, 19(7), 797818.[Taylor & Francis Online] [Google Scholar]) indicate that consumers prefer easy-to-access and easy-to-process information online.

Catering is the sub-industry of commercial industry, and its commercial tenant is a significant part of urban entity space, and a necessary and fundamental link of urban space units. Although there are many studies on spatial distribution features of urban catering tenants, research methods and data collection used are straightforward, and not enough attention has been paid on flow space and its spatial distribution pattern. With the popularization of network information technology, e-commerce is gradually changing the way of traditional marketing. It has increasingly become a significant factor that influences consumer decision-making; it may also influence catering provider's spatial distribution. E-commerce can be used to reflect the mobility (flow) of consumer dinning behavior without the limitation of spatial distance. Therefore, understanding spatial distribution pattern of urban catering industry and its E-WoM will be good explorations of central place theory and central flow theory.

2.4. Study area

Nanjing is one of the core cities of the Yangtze River Delta urban agglomeration, and its spatial pattern shows a typical multi-center hierarchical structure. According to the Nanjing Statistical Yearbook (2016 Nanjing Statistics Bureau (2016). Nanjing statistical year book-2016. Beijing: China Science and Technology Press. [Google Scholar]), as at the end of 2015, Nanjing has a total population of 8.23 million people. Nanjing has 27 business districts which include Xinjiekou, Hunan Road/Shanxi Road area, Confucius temple area, the Zhujiang Road area, the Taiping Road area, the Ruijin Road area, Nanjing University/Nanjing Normal University area, Caochangmen/Longjiang area, Xuanwu Lake Park area and the Sanpailou area. According to the Master Plan of Nanjing (2011–2020),11. ‘Master Plan of Nanjing (2011–2020)’ is the latest master plan of Nanjing prepared by the Nanjing Municipal Planning Bureau, and is used to determine the hierarchy of business centers in Nanjing.View all notes which provides the overall evaluation of a trading area according to its population scale and density, economic output, land price, infrastructure, etc., Xinjiekou is regarded as the highest grade trading area in the city, whereas the second highest grade areas include Hunan Road/Shanxi Road area, Confucius temple area, and the Zhujiang Road areas. Areas along the Taiping Road, the Ruijin Road, Nanjing University/Nanjing Normal University area, Caochangmen/Longjiang and Xuanwu Lake Park area are district business centers.

Established in April, 2003, Dianping.com is the earliest established third-party review website in China. The website establishes its pattern mainly with reference to Zagat Survey of the US. Registered members are able to post their comments freely on the website after their consumption in a restaurant, to provide objective and precise commentary information for potential customers. Till the third quarter of 2015, active users of dianping.com were over 200 million every month, the number of comments were over 100 million, the number of restaurants providers were over 20 million and it covered more than 250 cities in whole China, and had almost 20 billion monthly page views.22. http://www.dianping.com/aboutus.View all notes The website mainly includes information of services in the city such as food, leisure and recreation, and shopping, together with related activity sections such as coupons, groupons, and check-in deals. The section on catering has the most comment information, forming an enormous database of E-WoM that can influence decision-making in dining.

3. Methodology

Big data method is a full sample method based on data mining, and the data processing of relevant information recorded on the object, and the analysis of patterns is at a larger scale (Zhen et al., 2015 Zhen, F., Wang, B., & Wei, Z. (2015). The rise of the internet city in china: Production and consumption of internet information. Urban Studies, 52(13), 23132329.[Crossref], [Web of Science ®] [Google Scholar]). Applied to the studies of catering industries’ spatial pattern, on the one hand, big data can be used to understand urban dining space and distribution pattern in real time through restaurants location, reduce the limitation in slow data updating, and small sample restrictions of traditional research methods such as questionnaires or interviews. On the other hand, consumers can also use the evaluation score of each restaurant, to understand the overall development of the quality of the catering industry, compensating the missing information of restaurants such as quality, reputation and user preference in a traditional method.

3.1. Data collection

First, based on the internal standard of dianping.com, namely, ‘overall evaluation star level’, we removed comment information of 10,520 catering providers that haven't acquired evaluation star level. This is because (1) the restaurants may have low reputation, invalid information, and repeated registration; therefore, scored 0 on many indicators, making it hard to make calculations in the final model; (2) these restaurants have less attention and comments from consumers mainly because they are usually small in scale. As a result, 3645 valid catering providers remained. Both customer reviewers and restaurant data were collected. Such as reviewer's comments data include: Dish score (1–30 points), Atmosphere score (1–30 points), Service score (1–30 points), Per capita consumption, and Star-levels of overall comments (1–5 stars). The score was given by customers and calculated by the website itself. The restaurant data include the Number of web page visitors, Number of comments, Number of comments (credible), Number of comments by group buying, Number of interested people, Number of check-ins, Number of recommended dishes, Number of atmosphere comments, Number of special service (e.g. WiFi, parking space) comments, and Number of branches. These data were collected from the website. Second, we quantized text data, such as ‘average spending 50 RMB’, ‘three star’ in the overall score, and number of franchises. Information such as ‘average spending 50 RMB’ and ‘three star’ were calculated by the website itself (usually the average mean score) based on review scores. Data were then entered into EXCEL 2007, SPSS 20 and ArcGIS. Then, on the basis of catering providers’ detailed address, we built up a database of catering providers’ spatial location combined with Google Maps. No star evaluation does not represent poor service quality, but mainly indicates that lower attention was paid by customers on such catering providers. There are no secondary comments that were excluded from the data analysis. All data were collected from the website. The time range for the comments was collected from 2003 (since the website was created) to 30 August 2016.

3.2. Establishment of E-WoM evaluation index system

Based on the availability of the data and the previous studies (Baek et al., 2006 Baek, S. H., Ham, S., & Yang, I. S. (2006). A cross-cultural comparison of fast food restaurant selection criteria between Korean and Filipino college students. International Journal of Hospitality Management, 25(4), 683698.[Crossref] [Google Scholar]; Kincaid et al., 2010 Kincaid, C., Baloglu, S., Mao, Z., & Busser, J. (2010). What really brings them back?: The impact of tangible quality on affect and intention for casual dining restaurant patrons. International Journal of Contemporary Hospitality Management, 22(2), 209220(12).[Crossref], [Web of Science ®] [Google Scholar]; Kivela, 1997 Kivela, J. J. (1997). Restaurant marketing: Selection and segmentation in Hong Kong. International Journal of Contemporary Hospitality Management, 9(3), 116123.[Crossref] [Google Scholar]; Tao, Zhao, Yuan, & Tao, 2011 Tao, H., Zhao, Y., Yuan, X. Y., & Tao, P. (2011). The geographical position change and influence factors of Nanjing time-honored catering firms. World Regional Studies, 20(3), 145154. [In Chinese]. [Google Scholar]; Yüksel & Yüksel, 2003 Yüksel, A., & Yüksel, F. (2003). Measurement of tourist satisfaction with restaurant services: A segment-based approach. Journal of Vacation Marketing, 9(1), 5268.[Crossref] [Google Scholar]), we establish an E-WoM evaluation index system for classifying and screening data we obtained, from 6 aspects of comments on restaurants, namely, popularity, dish quality, atmosphere, service quality, scale and grade, and level (Table 1). There are 15 indicators at the second level: X1 is the total visit by customers, referring to the total number of visits customers have made to the website of a restaurant; X2 is the total number of comments, referring to the total number of comments made by customers; X3 is the default number of comments, referring to the reliable comments through web screening; X4 is the number of group comments, meaning the number of comments from those who bought group coupons from the website; X5 is customer favorite rate, referring to the number of customers who rate the restaurant as their favorite; X6 is check-ins, referring to the number of customers who visit the restaurants through the website of dianping.com; X7 is the overall rating of dishes by customers (scores 1–30); X8 is the total number of recommended dishes, referring to the total number of recommended dishes in a restaurant by customers (the website provides rating buttons for all dishes in the restaurant, and customers click on a particular dish to recommend it); X9 is the general atmosphere, referring to the score of general atmosphere in the restaurants by customers (1–30); X10 is the total number of atmosphere-type rating (the website provides voting buttons of atmosphere type, such as business gathering, friends gathering and family gathering, each time a customer clicks on a certain atmosphere and is supposed to be evaluated once); X11 is the total score of services, referring to customer rating of restaurant service (1–30); X12 is the total number of comments on special services, and thus the total number of comments by customers (the website provides rating buttons for special services, such as WIFI and parking, each time a customer clicks on a certain atmosphere and is supposed to be evaluated once); X13 is the average consumption per person; X14 is the number of branches, referring to the number of branches of this restaurant in Nanjing; X15 is the general star rating of customers, and thus the total rating of the restaurants by customers. We believe that E-WoM should not only include the direct comments (scores or star rating on the web by users) on food, service, environment, etc., but also include the attention from consumers. Generally speaking, restaurants with good E-WoM would attract more attention from consumers (such as websites browsed, comments made, dishes recommended and service recommended). Ideally, indexes should be set up according to theory. Due to the limited studies on evaluation of restaurants’ E-WoM score, the evaluation index was selected mainly based on data availability of the website while considering previously suggested elements of restaurants, such as dish quality, atmosphere and service quality (Baek et al., 2006 Baek, S. H., Ham, S., & Yang, I. S. (2006). A cross-cultural comparison of fast food restaurant selection criteria between Korean and Filipino college students. International Journal of Hospitality Management, 25(4), 683698.[Crossref] [Google Scholar]; Kincaid et al., 2010 Kincaid, C., Baloglu, S., Mao, Z., & Busser, J. (2010). What really brings them back?: The impact of tangible quality on affect and intention for casual dining restaurant patrons. International Journal of Contemporary Hospitality Management, 22(2), 209220(12).[Crossref], [Web of Science ®] [Google Scholar]; Kivela, 1997 Kivela, J. J. (1997). Restaurant marketing: Selection and segmentation in Hong Kong. International Journal of Contemporary Hospitality Management, 9(3), 116123.[Crossref] [Google Scholar]; Tao et al., 2011 Tao, H., Zhao, Y., Yuan, X. Y., & Tao, P. (2011). The geographical position change and influence factors of Nanjing time-honored catering firms. World Regional Studies, 20(3), 145154. [In Chinese]. [Google Scholar]; Yüksel & Yüksel, 2003 Yüksel, A., & Yüksel, F. (2003). Measurement of tourist satisfaction with restaurant services: A segment-based approach. Journal of Vacation Marketing, 9(1), 5268.[Crossref] [Google Scholar]). The 6 indexes at the first level were a combination of previous literature and current data on the website, and the 15 indexes at the second level were the data from the website and then put into each of the 6 different categories according to the nature of the index. A combination of these indicators can be used to measure restaurants’ E-WoM in a more comprehensive and complex way. Each of the 3645 catering providers include review data of those 15 indexes. The data size (number of reviews) is 54,675 (3645 × 15).

Table 1. E-WoM evaluation index system of catering industry in Nanjing urban area.

4. Results

4.1. Comprehensive assessment of the catering industry's E-WoM in Nanjing urban area

4.1.1. The E-WoM score of the catering industry in Nanjing varies widely

E-WoM scores vary widely in Nanjing. First, according to the E-WoM evaluation index system established earlier and the data of the 3645 catering providers from Daiping.com, a dataset was established in SPSS20. Then, KMO and Bartlett test were conducted. Results showed that the KMO value was 0.839 and Bartlett test of sphericity's P value was 0.000, which indicate a remarkable difference between correlation coefficient matrix and unit matrix. According to KMO metrics (generally, KMO value above 0.7 is suitable for Principal Component Analysis), the data was suitable for Principal Component Analysis. In order to illustrate interpretive degree of each principal component more clearly, the Varimax rotation was used in our analysis. From Table 2, we can see the eigenvalue of the four principal components were 5.594, 2.859, 1.992 and 1.035, all are greater than 1, and the cumulative contribution reached 76.531%. The first and the second components contribute a total of 56% to the E-WoM evaluation, while the third and fourth components only contribute 20%. Based on the principle of eigenvalue of principal component greater than 1 or cumulative contribution greater than 75%, it shows that the four principal components after Varimax rotation could well explain the E-WoM evaluation of catering industry in Nanjing urban area.

Table 2. Total variance explained.

According to Table 3, the highest loading of the first component is the number of recommended dishes, with a coefficient of 0.907 and this indicator suggests the popularity of dishes. The second component loading is the customers’ overall rating, with a coefficient of 0.932, showing residents’ overall impression of the restaurant. The third component is average spending per customer, with a coefficient of 0.722, showing the level of a restaurant; and the fourth component is the number of branches of the restaurants, with a coefficient of 0.981, showing the scale and brand of a restaurant.

Table 3. The factor loading matrix with orthogonal rotation.

By calculating the overall score, a comprehensive Principle Component Calculation model was obtained after the factor score of all restaurants (F1, F2, F3, F4) and taking the proportion of each principal component corresponding to the sum of the total eigenvalues of the extracted principal components as weights:where F is the score of Comprehensive Principle Component, and Ai is the weight coefficient of each principal component (variance contribution rate). It could be seen that E-WoM of the catering industry in Nanjing urban area varies widely, with a highest score reaching 119,572.64 while the lowest only 49.03. As can be seen from the value of each principal component, F1, F2 and F3 were all positive while F4 was negative, which indicates a strong correlation between restaurant's comment popularity, level evaluation, service quality and its overall degree of E-WoM. A weak correlation between restaurant's scale and grade and its overall degree of E-WoM was found, but it did not represent negative correlation. From another aspect, after testing and verifying data, it was seen that most catering providers with high overall score in E-WoM had high customer rates on overall star assessment on dianping.com, but catering providers with low E-WoM score did not necessary have low customer rates on overall star assessment on dianping.com, because of the low level and small amount of comments on these providers, which affected the evaluation of star level. Therefore, catering providers’ degree of E-WoM cannot be fully represented only by Dianping website's overall assessment star level.

4.1.2. The catering industry's E-WoM can be divided into four groups

We conducted k-means cluster analysis on Nanjing urban catering industry's composite degree of E-WoM. According to the overall degree of E-WoM distribution curve in Figure 1, four clusters were found. As such, the catering industry's degree of E-WoM can be divided into four groups. The first group consisted of five catering providers, whose E-WoM degree ranged between 119,573 and 84,672. The number of catering providers in this group was very small, and the slope of score was nearly 90°. This suggests that there are very limited restaurants of high E-WoM in Nanjing; and there is a large difference among restaurants’ E-WoM score, showing an uneven development. The second group has 38 providers, whose E-WoM degree ranged between 75,974 and 30,484. The number of catering providers in this group was larger than the first group, and the slope of score was nearly 90°. The third group consisted of 198 providers, whose E-WoM degree ranged between 29,706 and 9003. The number of providers in this group was obviously larger than the first two groups, and the slope of score was relatively flat, but still decreasing in large amplitude. The fourth group has 3404 providers, whose E-WoM degree ranged between 8998 and 47. The number of providers in this class was the largest while degree of E-WoM was the lowest. But the slope of score was quite gentle, and was of little decrease. This means that there is little difference among restaurants’ E-WoM score, showing an even development.

Figure 1. Distribution curve of catering providers’ E-WoM scores.

4.2. Spatial pattern of catering industry in Nanjing urban area

4.2.1. General spatial distribution patterns of catering providers

Using the detailed addresses of registered catering providers in Nanjing urban area on dianping.com, a spatial distribution figure of the 3645 catering providers was generated in ArcGIS software based on Google maps. It can be seen that catering providers are mainly located in the center of the urban area and four suburbs; however, a large difference exists between the number of catering providers in central urban area and suburbs. We obtained kernel density of catering providers in Nanjing urban area's distribution by using kernel density analysis tool in ArcGIS, setting a search radius of 500 m. It can be seen in Figure 2, the catering providers mainly spread around vital urban business circles. Among which, Xinjiekou has the highest distribution density with far larger cluster range compared to other areas while the distribution density and range of Hunan Road/Shanxi Road, Confucius temple and Zhujiang Road area takes the second place. The distribution density of Taiping Road, Nanjing University/Nanjing Normal University and Caochangmen/Longjiang area is also high, but the range is relatively small; only low-density and small-range clusters of catering providers can be found across Hexi Wanda Plaza, Zhongshan North Road/Yunnan Road, Ruijin Road and Gulou Park areas. Overall, following the scale and market proximity as the core indicators of the central place theory (Christaller, 1966 Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. [Google Scholar]; Krugman, 1991 Krugman, P. R. (1991). Increasing returns and economic geography.Southern Economic Journal, 99(3), 483500. [Google Scholar]), these distribution differences of restaurants in amount and density are highly related to the class of business circles (Zhang & Xu, 2009 Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. [Google Scholar]). Xinjiekou is the center of the city, providing a comprehensive function of business, finance, shopping, entertainment and hotels, and is regarded as the first tier urban business circle. Hunan Road, located north of the Xinjiekou circle, is the sub-business -circle, and mainly includes shopping and business. Confucius temple, located south of the Xinjiekou circle, is a commercial center with tourism, shopping and local snack shops mainly serving both tourists and residents. Zhujiang Road area is a shopping district for computers, and has its distinct features. Other business centers mainly provide shopping, business and entertainment, but mainly serve local residents. Meanwhile, Xinjiekou, Zhujiang Road, Taiping Road and Nanjing University/Nanjing Normal University areas have gathered contiguous development.

Figure 2. Distribution of catering providers in Nanjing City.

The results prove that although with the emergence of Internet platforms, the role of the central area has not disappeared, and the distribution pattern of urban catering providers reflects that the urban space still has a certain hierarchy. This is a typical reflection of central place theory (Christaller, 1966 Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. [Google Scholar]).

4.2.2. Spatial distribution features of catering providers’ E-WoM

4.2.2.1. General distribution features of catering providers

Generally speaking, the E-WoM distribution and density distribution of catering providers appear to be overlapping, with a tendency of ‘one core multi-heart’. In ArcGIS software, we set overall score of E-WoM as a calculating condition and set search radius of 500 m, and simulated the spatial distribution of E-WoM of catering providers in Nanjing urban area. As can be observed in Figure 3, E-WoM reaches the peak in Xinjiekou region which is the district with highest E-WoM group, whose range includes Hanzhong West Road, Chaotiangong, Zhongshan South Road and other regions. Hunan Road/Shanxi Road region has the second highest E-WoM peak and cluster scale, and becomes the second center. The Taiping Road area has the third highest E-WoM group while the Zhujiang Road and Confucius region have quite a large cluster of high E-WoM, whose score is lower and scale is also smaller compared to the first three regions. Xuanwu Lake Park, Jiangsu Road, Gulou Park, Nanjing University/Nanjing Normal University and Caochangmen/Longjiang region have high E-WoM, but their scale is quite small. The catering providers spreading near regional business center are not connected with large regions of high E-WoM like Hunan Road.

Figure 3. Distribution of catering providers’ E-WoM in Nanjing City.

We found that, different from the spatial distribution of the catering industry in Nanjing, which follows central place theory, the spatial distribution of E-WoM of the catering industry shows the features of central flow theory (Castells, 1989 Castells, M. (1989). The informational city: Information technology, economic restructuring, and the urban-regional process. Oxford: Blackwell. [Google Scholar], 1996 Castells, M. (1996). Rise of the network society: The information age: Economy, society and culture. Massachusetts: Blackwell Publishers. [Google Scholar]; Taylor, 2001 Taylor, P. J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181194.[Crossref], [Web of Science ®] [Google Scholar]; Taylor et al., 2008 Taylor, P. J., Evans, D. M., & Pain, K. (2008). Application of the interlocking network model to mega-city-regions: Measuring polycentricity within and beyond city-regions. Regional Studies, 42(8), 10791093.[Taylor & Francis Online], [Web of Science ®] [Google Scholar], 2010 Taylor, P. J., Hoyler, M., & Verbruggen, R. (2010). External urban relational process: Introducing central flow theory to complement central place theory. Urban Studies, 47(13), 28032818.[Crossref], [Web of Science ®] [Google Scholar]). First, there is a distinct hierarchal distribution of E-WoM, in which Xinjiekou forms the highest level, with Hunan Road, Taiping Road, Zhujiang Road, and Confucius temple areas gradually decreasing. Second, low-grade business circles do not necessarily mean low-grade E-WoM; for example, the level of business circles and the density of restaurants along Taiping Road are both lower than Confucius temple and Zhujiang Road area, but shows a higher grade in E-WoM. This may be because Taiping Road area is a famous leisure area which has many western restaurants, pubs, cafes, Japanese and Korean restaurants which are popular among young people. This area has a large population mobility and strong ability to gather popularity, and therefore acts as an important node in the spatial network of the catering industry in Nanjing. In addition, although the Nanjing catering industry shows contiguous distribution of levels (Figure 2), Figure 3 shows that the E-WoM in Xinjiekou, Hunan Road/Shanxi Road, Taiping Road, Confucius temple and Zhujiang Road areas is rather relatively independent and scattered in distribution, showing the characteristics of fragmentation.

To a large extent, the development of information technology, coupled with the fast development of Internet, breaks the traditional impact of distance on urban residents’ dining behavior. Information search engines and review platforms such as dianping.com facilitate residents’ choices of dining places from a wider area thereby making their dinning behavior more flexible. It also promotes the areas with ability to provide better or characteristic catering services and become the main areas of residential dining options, and the center of high-grade E-WoM, such as areas along Taiping Road.

4.2.2.2. Distribution features of different types of catering providers

Different types of catering providers have different degrees of consumer participation on E-WoM comments because they aim to differentiate consumer groups, and therefore should reflect a different spatial distribution feature of E-WoM. According to the business categories on dianping.com, two groups of catering providers were found: (1) mainstream cuisine which mainly includes Nanjing local cuisine, Huaiyang cuisine, Northern Jiangsu local cuisine, Sichuan cuisine, Cantonese cuisine, Hunan cuisine and North-east cuisine; (2) non-mainstream cuisine which mainly includes western food, Japanese and Korean cuisine, hotpot, snack and fast food, pastry and buffet. Data show that there are 1199 mainstream catering providers and 2446 non-mainstream catering providers in Nanjing urban area on dianping.com. Among the 241 providers that are of the top three E-WoM grades, the number of mainstream cuisine category is 2, 10 and 94, respectively, in the three grades, while for the non-mainstream cuisine category, 3, 28 and 104, respectively, showing approximately the same proportion in both categories. However, for the fourth grade, the amount of non-mainstream providers (2311) is far larger than the mainstream ones (1093), mainly because the non-mainstream catering includes more restaurants, which attracts a consumer group full of young people who often use dianping.com.

By analyzing the kernel density of 1199 mainstream catering providers’ E-WoM, setting a search radius of 500 m, we simulated the spatial distribution of mainstream catering providers in Nanjing urban area. From Figure 4, the amount of E-WoM peak of mainstream catering providers is smaller than that of all catering providers in Nanjing with a great difference. Xinjiekou is still the core region of E-WoM peak area; Hunan Road/Shanxi Road area becomes the second tier of high-grade E-WoM, while Confucius temple, Taiping Road and Xuanwu Lake areas become the third tier of E-WoM, while the Caochangmen/Longjian and Gulou Park areas become the regional peak area of E-WoM. From the distribution of mainstream cuisine E-WoM, the following features can be found: (1) a distinct hierarchical distribution of E-WoM, showing a similar distribution with the urban business district (Batten, 1995 Batten, D. (1995). Network cities: Creative urban agglomerations for the 21st century. Urban Studies, 32(2), 313327.[Crossref], [Web of Science ®] [Google Scholar]); (2) although Taiping Road and Xuanwu Lake area show a lower grade in business districts than Confucius temple area, the distribution of mainstream cuisine E-WoM shows the same grade, suggesting a strong popularity; (3) however, different from overall restaurant E-WoM distribution, the E-WoM distribution of mainstream cuisine is not independent or scattered, such that Xinjiekou, Confucius temple and Taiping Road areas have been joined contiguously, following a rule of traditional geographical gathering and distance reduction.

Figure 4. Distribution of mainstream catering providers’ E-WoM in Nanjing City.

Therefore, it is suggested that the E-WoM distribution of mainstream cuisine follows both rules of central place theory (Christaller, 1966 Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. [Google Scholar]) and central flow theory (Taylor 2001 Taylor, P. J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181194.[Crossref], [Web of Science ®] [Google Scholar]; Taylor et al., 2008 Taylor, P. J., Evans, D. M., & Pain, K. (2008). Application of the interlocking network model to mega-city-regions: Measuring polycentricity within and beyond city-regions. Regional Studies, 42(8), 10791093.[Taylor & Francis Online], [Web of Science ®] [Google Scholar], 2010 Taylor, P. J., Hoyler, M., & Verbruggen, R. (2010). External urban relational process: Introducing central flow theory to complement central place theory. Urban Studies, 47(13), 28032818.[Crossref], [Web of Science ®] [Google Scholar]). On the one hand, Internet platforms can help dinners to choose more flexibly (Bei et al., 2004 Bei, Lien-Ti, Chen, Etta Y. I., Rha, Jong-Youn, & Widdows, Richard (2004). Consumers’ online information search for a new restaurant for dining-out. Journal of Foodservice Business Research, 6(3), 1536.[Taylor & Francis Online] [Google Scholar]) and restaurants with a higher E-WoM could attract more visitors, thereby improving the E-WoM of the area. On the other hand, mainstream cuisines are usually chosen in formal occasions. High-level business districts, such as Xinjiekou area, have advantages due to their geographic location. They have more restaurants, higher mobility, better public facilities and transport accessibility, which make them popular areas for dinners. As a result, the high-level business district and the surrounding areas have high E-WoM restaurants, gathering a contiguous development.

By analyzing the kernel density of 2446 non-mainstream catering providers’ E-WoM value, setting a search radius of 500 m, we simulated the spatial distribution of non-mainstream catering providers in Nanjing urban area. From Figure 5, there is not much difference between the spatial distribution of E-WoM of non-mainstream catering providers and that of all catering providers in Nanjing urban area. Xinjiekou and Hunan Road/Shanxi Road areas form the first and second peak E-WoM of modern cuisine, Taiping Road and Zhujiang Road areas become the third peak E-WoM, while the Confucius temple and Chaotiangong areas are the fourth peak center. Other areas such as Nanjing University/Nanjing Normal University, Jiangsu Road and Xuanwu Lake are the regional peak E-WoM. The Zhujiang Road area shows a great difference between the E-WoM of overall restaurants and non-mainstream cuisines, demonstrating a higher score in non-mainstream cuisines’ E-WoM. This is the largest trading area for electronic products and services, and has a large number of young employed people and a shopping population, who prefer non-mainstream cuisines, which results into a higher grade in E-WoM for the area. In addition, different from the E-WoM distribution of mainstream cuisines, many consumers would choose snacks, western food, buffet, café in an informal occasion, and would prefer to choose restaurants with a good E-WoM via online platforms, resulting in more flexible and mobile dinning choices. This is reflected by the independent and scattered E-WoM distribution of non-mainstream restaurants.

Figure 5. Distribution of non-mainstream catering providers’ E-WoM in Nanjing City.

5. Conclusions

With the fast development of Internet and information technology, traditional central place theory has shifted to central flow theory, which focuses on ‘flow space.’ Although many studies have been conducted on the regional urban network and their flow patterns to explain interactions between cities (Naaman et al., 2012 Naaman, M., Zhang, A. X., & Brody, S. (2012). On the study of diurnal urban routines on Twitter. Sixth international AAAI conference on Weblogs and Social Media, Dublin. [Google Scholar]; Zhen et al., 2012 Zhen, F., Yu, Y., Wang, X., & Zhao, L. (2012). The spatial agglomeration characteristics of automotive service industry: A case study of Nanjing. Scientia Geographica Sinica, 32(10), 12001208. [In Chinese]. [Google Scholar]; Zook, 2001 Zook, M. A. (2001). Old hierarchies or new networks of centrality: The global geography of the Internet content market. American Behavioral Scientist, 44(10), 16791696.[Crossref], [Web of Science ®] [Google Scholar]; Zook & Graham, 2007 Zook, M. A., & Graham, M. (2007). Mapping digiplace: Geocoded internet data and the representation of place. Environment and Planning B: Planning and Design, 34(3), 466482.[Crossref], [Web of Science ®] [Google Scholar]), research on the spatial distribution of production factors within the city is still limited. Whether the features between regional organizations can still be found at the urban scale is not clear. In other words, how information technology has changed the flow space within the city and whether this type of change can be explained by central flow theory or central place theory needs further exploration.

In the catering industry, due to the fast development of information technology, especially the ICT platforms such as dianping.com, consumers’ dinning behavior has become more flexible and mobile rather than being limited by distance (Bei et al., 2004 Bei, Lien-Ti, Chen, Etta Y. I., Rha, Jong-Youn, & Widdows, Richard (2004). Consumers’ online information search for a new restaurant for dining-out. Journal of Foodservice Business Research, 6(3), 1536.[Taylor & Francis Online] [Google Scholar]; Qin et al., 2014 Qin, X., Zhen, F., Zhu, S. J., & Guang-Liang, X. I. (2014). Spatial pattern of catering industry in Nanjing urban area based on the degree of public praise from internet: A case study of dianping.com. Scientia Geographica Sinica, 34(7), 810817. [In Chinese]. [Google Scholar]). This trend of change has also brought new trends of research for urban geography, and the integration of Internet data and urban geographic spatial information has provided a new research content, method or research direction of urban geography (Graham & Shelton, 2013 Graham, M., & Shelton, T. (2013). Geography and the future of big data, big data and the future of geography. Dialogues in Human Geography, 3(3), 255261.[Crossref] [Google Scholar]).

This research uses urban catering space in Nanjing, China as a study object, which has significantly been influenced by new technology, and explores the application of central place theory and central flow theory in production elements within the city. After establishing a set of index of restaurants’ E-WoM from dianping.com (Nanjing), we calculated the complex score and ranking of catering providers’ E-WoM, and conducted kernel density analysis and a comprehensive evaluation on spatial distribution of the E-WoM of the urban catering industry and compared this ‘flow space’ with the general distribution patterns of catering providers (place space). The main conclusions are as follows:

(1)

The spatial distribution of restaurants’ density in Nanjing follows a traditional central place theory (Christaller, 1966 Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. [Google Scholar]; Krugman, 1991 Krugman, P. R. (1991). Increasing returns and economic geography.Southern Economic Journal, 99(3), 483500. [Google Scholar]; Zhang & Xu, 2009 Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. [Google Scholar]), showing a high overlapping of business districts. Xinjiekou is the center of the highest grade, Hunan Road/Shanxi Road, Confucius temple, and Zhujiang Road areas are the second highest grade, while Taiping Road, Nanjing University/Nanjing Normal University and Xuanwu Lake areas are the regional centers. Furthermore, the high-grade centers show the development trend of contiguous gathering. The results suggest the role of the central area has not disappeared, and the urban or regional space still has a certain hierarchy.

(2)

The spatial distribution of restaurants’ E-WoM in Nanjing follows the central flow theory (Castells, 1989 Castells, M. (1989). The informational city: Information technology, economic restructuring, and the urban-regional process. Oxford: Blackwell. [Google Scholar], 1996 Castells, M. (1996). Rise of the network society: The information age: Economy, society and culture. Massachusetts: Blackwell Publishers. [Google Scholar]; Taylor, 2001 Taylor, P. J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181194.[Crossref], [Web of Science ®] [Google Scholar]; Taylor et al., 2008 Taylor, P. J., Evans, D. M., & Pain, K. (2008). Application of the interlocking network model to mega-city-regions: Measuring polycentricity within and beyond city-regions. Regional Studies, 42(8), 10791093.[Taylor & Francis Online], [Web of Science ®] [Google Scholar], 2010 Taylor, P. J., Hoyler, M., & Verbruggen, R. (2010). External urban relational process: Introducing central flow theory to complement central place theory. Urban Studies, 47(13), 28032818.[Crossref], [Web of Science ®] [Google Scholar]). On the one hand, a clear hierarchy can be found in E-WoM distribution, but is not fully consistent with the level of urban business districts. For example, low-level business districts such as Taiping Road area are high grade in restaurant's E-WoM due to their high popularity. On the other hand, due to the rapid development and usage of the urban network information platforms, the influence of distance on residents’ consumption has weakened, and the dinning choice and activity space has been expanded, which makes the distribution of the restaurants’ E-WoM relatively scattered and independent.

(3)

Different characteristics can be found in different types of restaurants’ E-WoM distribution. Mainstream cuisine follows central place theory (clear hierarchy, contiguous development) but is inconsistent with urban business districts (such as Taiping Road and Xuanwu Lake areas), demonstrating some characteristics of central flow theory. The distribution of E-WoM of non-mainstream cuisine is similar to the E-WoM distribution of overall restaurants, showing a central flow system pattern. Therefore, our study finds that central place theory and central flow theory are still useful in explaining the spatial distribution of the catering industry and its E-WoM at the city scale.

The findings contribute to the following areas:

First, it explores the application of central place theory and central flow theory at the city scale, an aspect which was previously ignored by many researchers. It discusses the spatial distribution patterns of the catering industry considering network space, and examines whether the E-WoM reflects central flow theory with the influence of ICT. Previous research on central flow theory mainly considers the relationship of cities at the regional level. However, our research applied it to the urban level, and found that with the influence of information technology, urban internal dining shows a feature of central flow theory (clear hierarchy, grading is influenced by mobility and popularity, scattered and independent). However, a feature of traditional central place theory can still be found in the results such as ‘clear hierarchy and contiguous development’, mainly in the E-WoM of mainstream cuisine. It is obvious that information technology has provided more dinning flexibility and expanded the scope of dinning activity, and accelerated the dispersion of urban space elements. However, high-level urban business districts can still attract a variety of quality restaurants aggregation due to their advantages in scale, population mobility, facilities, and transport accessibility, such that the aggregational development of spatial elements is still evident. As such, a mixture of features of central flow theory and central place theory can be found at the city scale.

Second, the paper is a welcome addition of data analytical approach exploration into the hospitality industry, providing a more holistic understanding of catering industry in Nanjing. The integration of E-WoM scores from data mining and urban geographic spatial information (GIS method) has made it possible to examine the quality of restaurants at a larger scale.

The results have some practical implications as well. Studying urban service facility's quality and spatial distribution patterns is useful in decision-making for government's service facility planning. City governments should adopt various measures to optimize the space layout and minimize the unbalanced distribution of popular catering providers. For example, in order to balance the spatial distribution of different types of catering, priority can be given to non-mainstream cuisine in traditional business areas(such as Xinjiekou area) while mainstream cuisine can be encouraged in new business areas(such as Zhujiang Road) when planning catering facilities. In particular, the number and popularity of catering providers in the low-grade business centers may be increased, and the quality of catering needs to be improved.

Our research also has some limitations. Usage of massive data on websites to some extent reduces the influence of fake information cast on total sample, improving the accuracy of research data, and can generally reflect the spatial patterns of urban catering providers. But Internet users are mostly young people who tend to consume and comment on non-mainstream catering, which restrains the age range of the sample in this research. On the other hand, traditional central place theory lacks explanation on spatial distribution of urban catering industry. Establishing spatial structure pattern of urban catering industry based on E-WoM also needs comparison with other cities. Further research is still needed on the influencing factors, details and mechanism of information technology to the spatial distribution of urban catering industry.

Acknowledgments

The project is partly funded by Chinese National Nature Science Foundation (41571133; 41571146;41711530650), Central University Fundmental Research (2242017S10005; 2242015R30020) and China Postdoctoral Science Foundation (2017M611781). We would like to thank the three anonymous referees for their advices for this paper. Thanks also go to Dr. Felix Below for his help when preparing this paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. ‘Master Plan of Nanjing (2011–2020)’ is the latest master plan of Nanjing prepared by the Nanjing Municipal Planning Bureau, and is used to determine the hierarchy of business centers in Nanjing.

2. http://www.dianping.com/aboutus.

    References

  • Anderson, E. W. (1998). Customer satisfaction and word of mouth. Journal of Service Research, 1(1), 517. OpenURL University of Lincoln
  • Austin, S. B., Melly, S. J., Sanchez, B. N., Patel, A., Buka, S., & Gortmaker, S. L. (2005). Clustering of fast-food restaurants around schools: A novel application of spatial statistics to the study of food environments. American Journal of Public Health, 95(9), 15751581. OpenURL University of Lincoln
  • Baek, S. H., Ham, S., & Yang, I. S. (2006). A cross-cultural comparison of fast food restaurant selection criteria between Korean and Filipino college students. International Journal of Hospitality Management, 25(4), 683698. OpenURL University of Lincoln
  • Batten, D. (1995). Network cities: Creative urban agglomerations for the 21st century. Urban Studies, 32(2), 313327. OpenURL University of Lincoln
  • Bei, Lien-Ti, Chen, Etta Y. I., Rha, Jong-Youn, & Widdows, Richard (2004). Consumers’ online information search for a new restaurant for dining-out. Journal of Foodservice Business Research, 6(3), 1536. OpenURL University of Lincoln
  • Beldad, A., Jong, M. D., & Steehouder, M. (2010). How shall I trust the faceless and the intangible? A literature review on the antecedents of online trust. Computers in Human Behavior, 26(5), 857869. OpenURL University of Lincoln
  • Berger, J. (2014). Word of mouth and interpersonal communication: A review and directions for future research. Journal of Consumer Psychology, 24(4), 586607. OpenURL University of Lincoln
  • Bourlakis, M. A., & Weightman, P. W. H. (2004). Food supply chain management, Oxford: Blackwell Publishing. OpenURL University of Lincoln
  • Brown, J., Broderick, A. J., & Lee, N. (2007). Word of mouth communication within online communities: Conceptualizing the online social network. Journal of Interactive Marketing, 21(3), 220. OpenURL University of Lincoln
  • Brown, J. J., & Reingen, P. H. (1987). Social ties and word-of-mouth referral behavior. Journal of Consumer Research, 14(3), 350362. OpenURL University of Lincoln
  • Camagni, R., & Salone, C. (1993). Network urban structures in northern Italy: Elements for a theoretical framework. Urban Studies, 30(30), 10531064. OpenURL University of Lincoln
  • Castells, M. (1989). The informational city: Information technology, economic restructuring, and the urban-regional process. Oxford: Blackwell. OpenURL University of Lincoln
  • Castells, M. (1996). Rise of the network society: The information age: Economy, society and culture. Massachusetts: Blackwell Publishers. OpenURL University of Lincoln
  • Christaller, W. (1966). Central Places in Southern Germany. Englewood Cliffs, NJ: Prentice-Hall. OpenURL University of Lincoln
  • Davis, B., Lockwood, A., & Stone, S. (1998). Food and beverage management (3rd ed.). Oxford: Butterworth-Heinemann. OpenURL University of Lincoln
  • Eckardt, F. (2008). Media and urban space. In F. Eckardt (Eds.), Media and urban space: Understanding, investigating and approaching mediacity (pp. 79). Berlin: Frank & Timmer Gmbh. OpenURL University of Lincoln
  • Fusi, A., Guidetti, R., & Azapagic, A. (2016). Evaluation of environmental impacts in the catering sector: The case of pasta. Journal of Cleaner Production, 132(20), 146160. OpenURL University of Lincoln
  • Gottmann, J. (1961). Megalopolis, the urbanized northeastern seaboard of the United States. New York, NY: The Twentieth Century Fund. OpenURL University of Lincoln
  • Graham, M., & Shelton, T. (2013). Geography and the future of big data, big data and the future of geography. Dialogues in Human Geography, 3(3), 255261. OpenURL University of Lincoln
  • Gretzel, U., & Yoo, K. H. (2008). Use and impact of online travel reviews. Information and communication technologies in tourism, Enter 2008. Proceedings of the international conference in Innsbruck, (Vol. 26, pp. 3546). Vienna: DBLP. OpenURL University of Lincoln
  • Gwohshiung, T., Teng, M. H., Chen, J. J., & Opricovic, S. (2002). Multicriteria selection for a restaurant location in Taipei. International Journal of Hospitality Management, 21(2), 171187. OpenURL University of Lincoln
  • Hart, C., & Blackshaw, P. (2006). Internet INFERNO. Marketing Management, 15(1), 1825. Available from http://www.ehis.ebscohost.com [Accessed 26 January 2012]. OpenURL University of Lincoln
  • Hollenstein, L., & Purves, R. (2010). Exploring place through user-generated content: Using flickr to describe city cores. Journal of Spatial Information Science, 1(1), 2148. OpenURL University of Lincoln
  • Hu, Z. Y., & Zhang, Z. G. (2002). An analysis about the spatial distribution of hotels in urban area: Take Nanjing city as a case. Economic Geography, 22(1), 106110. [In Chinese] OpenURL University of Lincoln
  • Kang, C., Zhang, Y., Ma, X., & Liu, Y. (2012). Inferring properties and revealing geographical impacts of intercity mobile communication network of China using a subnet data set. International Journal of Geographical Information Science, 27 (3) 118. OpenURL University of Lincoln
  • Kincaid, C., Baloglu, S., Mao, Z., & Busser, J. (2010). What really brings them back?: The impact of tangible quality on affect and intention for casual dining restaurant patrons. International Journal of Contemporary Hospitality Management, 22(2), 209220(12). OpenURL University of Lincoln
  • King, R. A., Racherla, P., & Bush, V. D. (2014). What we know and don't know about online word-of-mouth: A review and synthesis of the literature. Journal of Interactive Marketing, 28(3), 167183. OpenURL University of Lincoln
  • Kivela, J. J. (1997). Restaurant marketing: Selection and segmentation in Hong Kong. International Journal of Contemporary Hospitality Management, 9(3), 116123. OpenURL University of Lincoln
  • Knaap, G. A. van der. (2002). Stedelijke bewegingsruimte, over veranderingen in stad en land. The Hague: Sdu Uitgevers. OpenURL University of Lincoln
  • Krings, G., Calabrese, F., & Ratti, C., & Blondel, V. D. (2009). Urban gravity: A model for inter-city telecommunication flows. Journal of Statistical Mechanics: Theory and Experiment, (7), 18. OpenURL University of Lincoln
  • Krugman, P. R. (1991). Increasing returns and economic geography.Southern Economic Journal, 99(3), 483500. OpenURL University of Lincoln
  • Liang, L. (2007). The distribution in space of urban catering and its factors: Xi'an as an example. Journal of Northwest University (Natural Science Edition), 37(6), 925930. [In Chinese]. OpenURL University of Lincoln
  • Litz, R. A., & Rajaguru, G. (2008). Does small store location matter? A test of three classic theories of retail location. Journal of Small Business & Entrepreneurship, 21(4), 477492. OpenURL University of Lincoln
  • Meijers, E. (2005). Polycentric urban regions and the quest for synergy: Is a network of cities more than the sum of the parts? Urban Studies, 42(42), 765781. OpenURL University of Lincoln
  • Melaniphy, J. C. (1992). Restaurant and fast-food site selection. New York, NY: Wiley. OpenURL University of Lincoln
  • Muller, C. C., & Inman, C. (1994). The geodemographics of restaurant development. Cornell Hospitality Quarterly, 35(3), 8895. OpenURL University of Lincoln
  • Naaman, M., Zhang, A. X., & Brody, S. (2012). On the study of diurnal urban routines on Twitter. Sixth international AAAI conference on Weblogs and Social Media, Dublin. OpenURL University of Lincoln
  • Nanjing Statistics Bureau (2016). Nanjing statistical year book-2016. Beijing: China Science and Technology Press. OpenURL University of Lincoln
  • Ohlin, B. (1993). 1933 and 1977–some expansion policy problems in cases of unbalanced domestic and international economic relations. American Economic Review, 83(6), 1017. OpenURL University of Lincoln
  • Qin, X., Zhen, F., Zhu, S. J., & Guang-Liang, X. I. (2014). Spatial pattern of catering industry in Nanjing urban area based on the degree of public praise from internet: A case study of dianping.com. Scientia Geographica Sinica, 34(7), 810817. [In Chinese]. OpenURL University of Lincoln
  • Schaefer, A. D., Luke, R. H., & Green, J. (1996). Attitudes of restaurant site selection executives toward various people magnets. Journal of Restaurant & Foodservice Marketing, 1(3), 114. OpenURL University of Lincoln
  • Schwanen, T., Dijst, M., & Kwan, M. P. (2006). Introduction-the internet, changing mobilities, and urban dynamics. Urban Geography, 27(7), 585589. OpenURL University of Lincoln
  • Senecal, S., & Nantel, J. (2004). The influence of online product recommendations on consumers online choices. Journal of Retailing, 80, 159169. OpenURL University of Lincoln
  • Shu, S., Wang, R., Sun, Y., Liu, J., & Xiao, L. (2012). Spatial distribution of urban catering industry and its influenced factors: A case study of Xiamen City. Tropical Geography, 32(2), 134140. OpenURL University of Lincoln
  • Smith, S. L. J. (1985). Location patterns of urban restaurants. Annals of Tourism Research, 12(4), 581602. OpenURL University of Lincoln
  • Sparks, B. A., & Browning, V. (2010). Complaining in cyberspace: The motives and forms of hotel guests’ complaints online. Journal of Hospitality Marketing & Management, 19(7), 797818. OpenURL University of Lincoln
  • Tao, H., Zhao, Y., Yuan, X. Y., & Tao, P. (2011). The geographical position change and influence factors of Nanjing time-honored catering firms. World Regional Studies, 20(3), 145154. [In Chinese]. OpenURL University of Lincoln
  • Taylor, P. J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181194. OpenURL University of Lincoln
  • Taylor, P. J., Evans, D. M., & Pain, K. (2008). Application of the interlocking network model to mega-city-regions: Measuring polycentricity within and beyond city-regions. Regional Studies, 42(8), 10791093. OpenURL University of Lincoln
  • Taylor, P. J., Hoyler, M., & Verbruggen, R. (2010). External urban relational process: Introducing central flow theory to complement central place theory. Urban Studies, 47(13), 28032818. OpenURL University of Lincoln
  • Teller, C., & Reutterer, T. (2008). The evolving concept of retail attractiveness: What makes retail agglomerations attractive when customers shop at them? Journal of Retailing & Consumer Services, 15(3), 127143. OpenURL University of Lincoln
  • Timor, M., & Sipahi, S. (2005). Fast-food restaurant site selection factor evaluation by the Analytical Hierarchy Process. The Business Review Cambridge, 4(1), 161167. OpenURL University of Lincoln
  • Yoon, S. (2009). The effects of electronic word-of-mouth systems (EWOMS) on the acceptance of recommendation (Dissertation Abstracts International Section A, 69, 7-A). PsycINFO, EBSCOhost. OpenURL University of Lincoln
  • Yüksel, A., & Yüksel, F. (2003). Measurement of tourist satisfaction with restaurant services: A segment-based approach. Journal of Vacation Marketing, 9(1), 5268. OpenURL University of Lincoln
  • Zhang, X., & Xu, Y. L. (2009). Study on the distribution in space of urban caterings and its influencing factors: A case study of Nanjing. Tropical Geography, 29(4), 134140. [In Chinese]. OpenURL University of Lincoln
  • Zhen, F., Liu, X. X., & Liu, H. (2007). Regional urban network influenced by information technology: New directions of urban studies. Human Geography, 22(2), 7680. [In Chinese]. OpenURL University of Lincoln
  • Zhen, F., Wang, B., & Wei, Z. (2015). The rise of the internet city in china: Production and consumption of internet information. Urban Studies, 52(13), 23132329. OpenURL University of Lincoln
  • Zhen, F., Yu, Y., Wang, X., & Zhao, L. (2012). The spatial agglomeration characteristics of automotive service industry: A case study of Nanjing. Scientia Geographica Sinica, 32(10), 12001208. [In Chinese]. OpenURL University of Lincoln
  • Zhou, K. H., Zhen, F., Yu, Y., & Jiang, Y. H. (2010). A research on the processes and patterns of spatial agglomeration of financial services in urban central area: A case study of Kuiwen district, Weifang city. Human Geography, 25(6), 6267. [In Chinese]. OpenURL University of Lincoln
  • Zhu, F., & Zhang, X. (2012). Impact of online consumer reviews on sales: The moderating role of product and consumer characteristics. Journal of Marketing, 74(2), 133148. OpenURL University of Lincoln
  • Zook, M. A. (2001). Old hierarchies or new networks of centrality: The global geography of the Internet content market. American Behavioral Scientist, 44(10), 16791696. OpenURL University of Lincoln
  • Zook, M. A., & Graham, M. (2007). Mapping digiplace: Geocoded internet data and the representation of place. Environment and Planning B: Planning and Design, 34(3), 466482. OpenURL University of Lincoln

Additional information

Author information

Feifei Xu

Feifei Xu, PhD, is a professor in the School of Humanities, Southeast University, China. Her main research interests include tourism in protected areas, cross-cultural issues in tourism, as well as e-tourism.

Feng Zhen

Feng Zhen, PhD, is a professor in the School of Architecture and Urban Planning, Nanjing University, China. His main research interests include the impacts of Information communication and technology (ICT) on urban space and smart city theory.

Xiao Qin

Xiao Qin, PhD, is an assistant researcher in the School of Architecture and Urban Planning, Nanjing University, China. His main research interests include the application of big data in urban planning.

Xia Wang

Xia Wang, PhD, is an associate professor in the School of Geographic and Oceanographic Sciences, Nanjing University, China. Her main research interests include the tourism geography and smart tourism.

Fang Wang

Fang Wang, PhD, is a professor in the College of Architecture and Landscape Architecture, Peaking University, China. Her main research interests include the cultural landscape and Geo-Architecture.

Funding

This work was supported by Fundamental Research Funds for Central University [grant number 2242015R30020], [grant number 2242017S10005]; Chinese National Nature Science Foundation [grant number 41571133], [grant number 41571146], [grant number 41711530650]; China Postdoctoral Science Foundation [grant number 2017M611781].