TL;DR
This paper leverages geotagged photos from Flickr to map and analyze the spatial characteristics of inbound tourist flows in China from 2011-2017. By employing Markov chains, frequent-pattern mining, and community detection, it reveals key travel patterns, identifies major tourist hubs forming a "double-triangle" framework, and categorizes cities into nine distinct tourism groups, offering valuable insights for tourism development.
What Problem Does This Paper Address?
Understanding tourist movement is crucial for effective tourism management, marketing, and product development. However, obtaining granular, city-level data on tourist flows, especially in developing countries, has been a significant challenge. Traditional methods like questionnaires are time-consuming, provide limited data samples, and can lead to biased results due to scale dependency. This paper aims to overcome these limitations by utilizing the wealth of data available from geotagged photos on social media platforms to provide a more comprehensive and accurate picture of inbound tourist behavior in China.
Methods
The study employs a multi-faceted approach to analyze inbound tourist flows:
- Data Acquisition: Geotagged photos and their associated metadata (user ID, time, location) were collected from Flickr's API for China (excluding Taiwan, Hong Kong, and Macao) spanning from January 2011 to December 2017.
- Markov Chain Model: This statistical model is used to describe the probability of a tourist transferring from one city to another. It assumes that the probability of moving to the next city depends only on the current city, not the entire travel history.
- Frequent-Pattern Mining (Apriori Algorithm): This technique is applied to identify common travel routes or "patterns" among tourists. It looks for sequences of visited cities that occur frequently within the dataset.
- Community Detection (Combo Algorithm): This algorithm is used to group cities into clusters based on the strength of their interconnections (tourist flow intensity). The goal is to identify communities where tourist flows are dense within the group and sparse between groups, revealing natural tourism regions.
Datasets
- Geotagged Photos from Flickr: The primary dataset consists of 282,787 geotagged photo records from Flickr, which were filtered down to 80,278 records from 3680 unique tourists after excluding Chinese users and those with incomplete data. This dataset covers the period from January 2011 to December 2017.
Key Insights and Ideas
- The "Double-Triangle" Framework: The strongest inbound tourist flows are concentrated between Beijing and Shanghai, forming a crucial backbone. When combined with Xi'an and Guilin, these cities create a "double-triangle" framework that significantly influences overall inbound tourist movement.
- Emerging Destinations Gaining Traction: Travel to emerging destination cities in Central and Western China is becoming increasingly popular, indicating a diversification of tourist interests beyond traditional hotspots.
- Geotagged Photos as a Rich Data Source: The study demonstrates the power of leveraging user-generated content from social media platforms like Flickr for large-scale spatial analysis of human behavior, overcoming limitations of traditional data sources.
- Community Structure Reflects Regional Dynamics: The identified nine tourism groups highlight how inbound tourist flows are influenced not only by natural and cultural attractions but also by regional economic development and existing infrastructure.
Key Results
- Strongest Flow: The strongest inbound tourist flow is between Beijing and Shanghai.
- "Double-Triangle" Framework: Beijing, Shanghai, Xi'an, and Guilin form a "double-triangle" inbound tourist flow framework.
- Emerging Itineraries: Travel to emerging destination cities in Central and Western China is becoming more frequent.
- Nine Tourist City Groups: Based on flow intensity, inbound tourist destination cities were divided into nine distinct groups.
- Popular Travel Patterns: The most frequent two-city travel patterns include Beijing→Shanghai and Shanghai→Beijing. For three-city patterns, Shanghai→Xi'an→Beijing emerged as a prominent route, indicating these cities as key hubs.
- Yangtze River Cruise Impact: The study notes the significant effect of Yangtze River cruise tourism, with routes like Chongqing-Yichang to Beijing or Shanghai showing up in frequent travel patterns.
Limitations and Future Work
- Country-Specific Analysis: The study analyzed inbound tourist data as a whole. However, travel behaviors can vary significantly among tourists from different countries and regions. Future research should analyze these differences in detail.
- Influence of External Factors: While the study focuses on spatial flows, it acknowledges that transportation, marketing, economic development, and public services significantly impact inbound tourism. Future research should quantitatively analyze the influence of these factors.
- Weak Connections in Hainan: The paper notes that cities in Hainan Island show weak connections, which might be related to vacation travel preferences of international tourists. This area warrants further exploration.
- Individualized Demands: The study highlights that while strong flows are well-understood, the individualized demands of "weak tourist flows" could drive innovation in tourism products and encourage regional cooperation. Further investigation into these less common but potentially innovative demands is suggested.
Summary updated