Diet is not only a basic material element to meet human physiological needs, but it is also an important carrier of human cultural elements. The cultural phenomena related to the demand, production, and consumption of food were investigated in food culture studies [1
]. Food culture has the characteristics of multi-disciplinary research, and the study of food culture from the perspective of geography emphasizes the human–environment relationship [3
] and the geographical differentiation law reflected by the preference of food culture [4
]. Geographical regionalization and mapping of food culture are important analytical and presentation methods to reveal regional differences of food culture, which are significant for mining local food culture resources and understanding local food culture characteristics [6
]. At the same time, the regional structural characteristics presented by the food culture regionalization provide references for studying human–environment relationship in the geography of food culture [7
Research on regional food culture is a trending topic in geography and related disciplines. Research topic includes two aspects, namely, what influences regional food cultures and what regional food cultures have influence on. The former focuses on how economy, politics, history, religion, and geographical environment affect regional food culture, whereas the latter focuses on the impact of regional food culture on people’s health, tourism, and other aspects, namely, impact of regional food cultures on the formation of sense of place and the construction of local cultural brands and symbols. In terms of the influence of economy on food culture, certain scholars reported that fast food chains in developed regions will be the first to enter and then affect the local food culture [8
]. Several researchers also claimed that groups with different incomes will produce various food consumption cultures [10
]. In the research on the relationship between regional food culture and politics, Tellstrom et al. studied how governments interpret and present their national image by shaping local food culture [11
]; that is, how the differences in political culture affect the food culture and daily food consumption patterns of people in different regions [12
]. Historical and cultural customs and population migration in the history also profoundly impact local food cultures [14
]. The relationship between religions and regional food culture is also a hot topic; for example, religious beliefs affect structures of food cultures, and the degree of attention to the green food differs between religious and nonreligious people [15
In studies on geographical environment and food culture, climate, food materials and their proximity to coastal areas are the core influencing factors for the formation of regional food culture [18
]. These studies discussed food culture as affected objects, that is, which factors affect the production and reproduction of food cultures. In the studies of food cultures, the relationship between regional food culture and health attracted the attention of scholars as the major influencing factor. For example, certain studies suggested that the food cultures of different regions always tend to develop in the direction that is beneficial to people’s health. The food cultures of tropical regions help people keep cool, whereas those of cold regions help people stay warm [21
]. Certain studies also suggested that the existing food culture in certain places is inconducive to people’s health [23
]. Another widely followed research topic is how regional food culture affects the development of tourism. Several studies reported that people’s pursuit of local characteristic cuisine stimulated the interest of tourism developers and tourists in food tourism. In terms of academic research, food tourism is also an expanding field [25
]. For example, Robinson et al. conducted a data-driven empirical study on gastronomic tourism among tourists [26
]. In recent years, several scholars also focused on the production of cross-regional food culture and the construction of local sense [27
Most studies regarded food as a culture and emphasized regional characteristics as inherent geographical attributes of food cultures. These studies also highlighted the human–environment relationship reflected in food cultures. However, the regional characteristics of people’s food preference differences emphasized by geography are rarely mentioned in the literature. Only several studies are available on the establishment of boundaries of food culture and homogeneous-culture regions by geographical regionalization method when the same region involves multiple food cultures. China puts importance on food culture. However, it has vast territory, diverse food culture, and evident regional differences in food culture. Hence, conducting geographical regionalization based on various food cultures in the Chinese context is necessary to have a deep understanding of the characteristics of regional food cultures.
Cultural regionalization is the basis for deeply mining regional cultural characteristics and scientifically cognizing regional differences in culture [29
]. It is an important way to extract and understand cultural regions. Cultural region is one of five main themes in traditional cultural geography research and the others are cultural landscape, ecology, diffusion and integration [30
]. As an important cultural element in China, the food culture represented by Chinese cuisines has a profound influence on the regional structure of Chinese culture [30
]. Existing researches are mainly based on cultural elements to extract culture regions, such as dialects and history [32
]. However, regional culture is a process of dynamic change. Some cultural phenomena show a long-term slow and gentle dynamic process of change (such as dialects and history), while others show a faster and sharper process of dynamic change [33
]. Food represents a cultural element of rapid change [36
]. This paper is based on the study of food culture regionalization of restaurant POI data which owns nearly the whole restaurant samples in China. It is of great significance for understanding the regional characteristics of food culture and enhancing the identity of regional culture [29
In terms of research methods, geographical regionalization is usually adopted in the following ways: semi-manual cartography method based on experience and data [38
], traditional GIS overlay analysis and cartographic synthesis method [40
], and clustering method based on machine learning [42
]. The semi-manual cartography method has low quantification degree and large workload. Although methods based on overlay analysis and cartographic synthesis have a high degree of automation, handling the weight distribution of geographical objects with complex features is difficult. At present, the multisource clustering method based on machine learning has been introduced into the research of geographical regionalization and extended into the multisource clustering method with spatial constraint. In addition to its ability to handle the complex geographical regionalization with multi-source characteristics, the main advantage of this method is that it is a clustering method based on object model. The multisource clustering method with spatial constraints is adopted to conduct geographical regionalization of food cultures with multiple characteristics, which not only give play to the advantages of the method but also solve two problems, namely, the complex features of cultural phenomenon and the difficulty of using traditional method for a reasonable regionalization result.
Density based hotspot detection is another appropriate way to discover the spatial quantitative characteristics due to it can identify the spatial distribution pattern of objects or elements [44
]. By the way of hotspot detection, some researchers have studied on spatial patterns of points with different semantics. Typical hotspot detecting algorithms include DBSCAN, OPTICS, Kernel Density Estimation (KDE), etc. The analysis results of DBSCAN and OPTICS are density-connected objects and tell us which objects with high density [45
]. By contrast, the analysis result of KDE is local maximums based on the density field, thus tell us which area with high density [43
]. However, these methods are difficult to extract the specific locations with the local maximum density, this paper proposes a density-based hot spot detection method to achieve this goal, so that better represents spatial quantitative characteristics of restaurants.
This study takes the entire China as research area, adopts the restaurants of the eight most famous great cuisines as the main data source, and takes prefecture-level cities as the main research units. It uses the preference index of each prefecture-level cities for each cuisine as the regional analysis variable and adopts regionalization method with spatial constraints for geographical regionalization of food cultures. The principle is to maintain the similarity of food preference structure within the same area and maximize the difference in different regions. Finally, China is divided into several continuous food culture regions. Thus, the quantification of the regional structure of Chinese food culture based on the eight traditional cuisines is conducive to deep cognition of the spatial differentiation rules of Chinese food culture and excavation of local characteristic food culture.
2. Study Area and Data Description
The cultural background of Chinese food is different from that of Western countries such as Europe and the United States. The majority of Chinese people’s diet is mainly Chinese food. Although Chinese food is made up of various regional cuisines, it includes not only the eight major cuisines. After a long period of evolution and its own system, the eight major cuisines with distinctive local flavored characteristics, are widely recognized by the society and the most influential local cuisines in China. In China, food preferences vary greatly from region to region, and have always been seen as a cultural symbol which is used to distinguish cultural differences between people in different regions. This provides rationality for treating food preferences as a cultural phenomenon, as well as the possibility of regionalization based on food culture. From the perspective of cultural influence, the eight major cuisines can be regarded as the representation of food preference. The concentration of restaurants with eight major cuisines indicates the concentration of their audiences, and the proportion of different cuisines in the areas indicates the composition of food preferences in the areas. Thus, Clustering and regionalization based on eight major cuisines can represent the regional structure of food culture preference.
The research area is Mainland China. Taiwan, Hong Kong, and Macao are excluded due to the lack of data on catering facilities in these regions. The research area comprises 31 secondary administrative regions (provinces, autonomous regions, and municipalities). In the regional classification and geographical regionalization of food cultures, the basic geographical unit adopted in this study is prefecture-level city, which is the third administrative regionalization after the province or autonomous region. Prefecture-level cities are adopted as the analysis unit because their population reached a certain scale, and the number of catering service facilities can guarantee the diversity to a certain extent to avoid the situation where the number of certain cuisines is zero in certain analysis units. Assuming that lower-level county is taken as the analysis unit, not all counties would have restaurants with the eight major cuisines. Because there are 2844 county-level regions in the country, and the total number of restaurants with certain types of the eight major cuisines is not larger than 5000. The whole research area is shown in Figure 1
. The yellow highlights in Figure 1
are the provinces where the eight major cuisines are located and the birthplace of their respective cuisine culture.
As the representative of Chinese food culture, the eight major cuisines have a long history of inheritance. These specialty restaurants can now be found in most Chinese cities, with some expanding rapidly and others slowly. For example, Sichuan and Hunan cuisines have the fastest diffusion, whereas Jiangsu and Shandong cuisines have the slowest (Table 1
). The data was extracted from the online map service of AMap in 2018, China’s most famous online map service provider. In AMap map, the POI of catering type contains the classification field of specific catering type, which contains the classification value of traditional eight cuisines. Eight major cuisines and their geographical location information can be extracted using this field. Table 1
lists the names of the eight major cuisines, major origins, total number of restaurants nationwide, and percentage of restaurants in each cuisine.
5. Conclusions and Future Direction
The study of food cultures from the perspective of geography emphasizes the human–environment relationship and the law of regional differentiation reflected by food cultures. Various characteristic cuisines are the carrier of local food cultures and the symbol of regional food cultures. The formation of various cuisines (such as the eight major cuisines in this study) is the objective presentation of regional differences in food cultures. Starting from the research paradigm of cultural geography, this study analyzes the spatial diffusion trend and regional differentiation pattern of food culture reflected by the eight most distinctive cuisines in China from the perspectives of quantitative and regional characteristics. The spread of the eight traditional Chinese cuisines is primarily manifested in two aspects; that is, spatial distribution and spatial hierarchy. Regional differentiation is primarily reflected in the formation of food culture regions and the emergence of new cultural boundaries.
From the perspective of the quantitative characteristics of these cuisines, the farther south the regional cuisines are, the greater the spatial diffusion scope of cuisine food culture is in terms of spatial distribution. By contrast, the farther north the cuisines are, the more limited the scope is. Among these cuisines, Sichuan cuisine presents the most vigorous space diffusion, whereas Shandong cuisine presents the weakest space diffusion. For most types of cuisines, urban agglomerations, such as the Yangtze River Delta, the Pearl River Delta, and the Beijing–Tianjin–Hebei Regions, do not form the hotspots of absolute superiority due to their population aggregation, developed economy, and large population mobility. Instead, these agglomerations show a trend of outward diffusion from the origin of each cuisine. In terms of hierarchical structure, the hotspots of different levels of Shandong, Jiangsu, and Anhui cuisines are uneven, and polarization in these hotspots is evident. Other southern cuisines have a relatively balanced hotspots distribution in five grades. To sum up, regardless of spatial distribution or hierarchical structure, the overall trend of north–south differentiation is shown. The former is reflected in the diffusion range, whereas the latter is reflected in the equalization of hotspots at different levels.
This study uses prefecture-level cities as the basic research unit and takes the proportion of each cuisine as the variable to make comprehensive regionalization. The geographical regionalization results of the preference of the eight Chinese cuisines are actually the construction process of food culture regions. Each food culture region not only maintains the spatial continuity but also ensures the similarity of people’s preferences for these cuisines in the cultural region. Cultural districts ensure the diversity of their preferences for these cuisines. Cultural regions and their boundaries are in dynamic change, deeply influenced by the daily production and reproduction of cuisine culture. This study constructs new cultural regions using the restaurant data for the understanding of the characteristics of food culture in different food culture regions. This study also suggests that cultural boundaries have not been disappeared because of the development of modern transportation technology and the Internet, which give a reverse response to the concern on geographical boundaries will be gone according to “The Exaggerated Death of Geography”. In addition, the cultural boundary is a core issue in the field of cultural geography. This paper also provides a feasible way of thinking for the cultural boundary of quantitative extraction.
Although this paper has analyzed the spatial distribution pattern and regional structure characteristics of Chinese food culture based on restaurant POI data, there is still some work that needs to be further involved. The target of quantitative analysis of spatial structure is mechanism analysis. In future, on the one hand, we will further study how the culture region structure formed, and main factors which affected. On the other hand, only one year’s restaurant data is used to regionalization in this paper, the future study will focus on different years dynamic change characteristics of the regional structure, though there is still a challenge on the acquisition of restaurant POI data for long time series in the whole of China.