TL;DR
This paper uses restaurant Point of Interest (POI) data from China to map and understand the regional structure of food culture, moving beyond simple cuisine popularity to analyze regional preferences. By developing a density-based hotspot detection method and a regionalization approach, they reveal distinct spatial patterns and cultural boundaries shaped by cuisine preferences.
What Problem Does This Paper Address?
While food is recognized as a significant cultural element, its regional differentiation and the spatial structure of food culture haven't been thoroughly explored from a geographical perspective. Existing research often focuses on factors influencing food culture (like environment or economy) or its impact (on health or tourism), but lacks a deep dive into how cuisine preferences themselves create distinct regional patterns. This paper aims to fill that gap by using restaurant data to understand the spatial organization of food culture in China, a country with vast territory and diverse culinary traditions.
Methods
The study employs a two-pronged approach:
- Density-Based Hotspot Detection: To identify areas with a high concentration of restaurants for each of the eight major Chinese cuisines. This method, an adaptation of kernel density estimation, helps visualize the spatial quantitative characteristics of each cuisine.
- Geographical Regionalization: After identifying hotspots, the researchers use a machine learning-based heuristic method to classify prefecture-level cities based on the proportion of each of the eight cuisines. This moves beyond raw counts to understand the relative preference for different cuisines within a region. This method incorporates spatial constraints to ensure regional coherence and similarity within identified zones.
Datasets
The primary dataset used is restaurant Point of Interest (POI) data from AMap (China's most popular online map service) for 2018. This dataset contains approximately 7.5 million restaurants, with specific labels for the eight major Chinese cuisines. The study area is Mainland China, excluding Taiwan, Hong Kong, and Macao due to data limitations.
Key Insights and Ideas
- Cuisine Diffusion Varies: Some cuisines, like Sichuan and Hunan, show strong diffusion across the country, while others, like Shandong and Jiangsu, are more geographically concentrated.
- Hotspots Aren't Always in Origin Regions: Surprisingly, some cuisines' primary hotspots (areas of high restaurant density) are found outside their traditional cultural hearths, often in economically developed areas like Shanghai or Guangzhou.
- Regionalization Reveals Deeper Structure: By analyzing the proportion of cuisines rather than just raw numbers, the study uncovers distinct regional food culture zones that reflect nuanced local preferences.
- Cultural Boundaries Persist: Despite modern transportation and the internet, distinct regional food cultures and their boundaries remain, suggesting that geography still plays a significant role in shaping culinary identity.
Key Results
- The study identified spatial distribution patterns and hierarchical structures for eight major Chinese cuisines, revealing varying degrees of spatial diffusion and concentration.
- A regionalization map of China was generated, dividing the country into eight distinct food culture regions based on cuisine preference. These regions reflect similarities in cuisine proportions within their boundaries.
- The research highlights that southern cuisines generally exhibit stronger spatial diffusion than northern cuisines. Sichuan cuisine was found to have the strongest diffusion, while Shandong had the weakest.
- The study found that while some cuisines have stable proportions across many cities, others show significant instability, indicating dynamic regional preferences.
Limitations and Future Work
- Data Limitations: Taiwan, Hong Kong, and Macao were excluded due to a lack of catering facility data.
- Temporal Scope: The analysis is based on restaurant data from a single year (2018). The authors acknowledge that future work should focus on analyzing dynamic changes in regional structures over multiple years, which presents a challenge due to data acquisition.
- Mechanism Analysis: While the study identifies spatial patterns and regional structures, future research should delve deeper into the underlying mechanisms and factors that shape these formations.
Summary updated