2:15 PM - 2:30 PM
[HTT14-08] Network Analysis on the Evolution of Tourism Destination Network: A Case Study of Inbound Tourism to Japan
Keywords:Network analysis, International tourism, Spatio-temporal pattern
Purpose of the study:
Several studies have explored geographical networks from a spatio-temporal perspective, including transportation, urban systems, and goods supply networks. However, the evolution of tourism destination networks has received less attention, particularly on a regional or national scale due to a lack of time-series flow data. The tourism destination network is dynamic and continuously changing, and understanding its evolution is crucial in addressing a range of challenges, such as overtourism and regional revitalization. This research aims to use network analysis to estimate the spatio-temporal evolution of the tourism destination network in Japan. The study focuses on inbound tourism in the 2000s, specifically during the target period of analysis, as inbound tourism during that period had a significant impact due to the implementation of the Visit Japan Campaign policy. The research investigates three main questions:
RQ1: Will the tourism destination network become more centralized or decentralized?
RQ2: Will more clusters emerge over time?
RQ3: What are the influential factors shaping the network?
Data and Methodology:
The study considers the itinerary of package tours to construct the networks between destinations. The target market is Taiwan, which is one of the major inbound markets to Japan. The data was obtained from the "Wayback Machine," a long-term website archive. Historical package tours sold by a major travel agent during the period of 2002-2008 were collected. After manual data cleaning, an adjacency matrix of 212 city-level destinations in Japan was created.
This study adopts network analysis, a quantitative methodology rooted in graph theory that has been utilized in geography since the 1960s, to examine the relationships between actors. In this study, the destinations are treated as actors, and the touring routes are regarded as links. The analysis consists of three parts: the centralization index, which reflects the global character of the network; community detection, which reveals the local character of the network; and Quadratic Assignment Procedure (QAP), which identifies the influential factors that shape the network.
Results and discussions
The study reveals three main findings. Firstly, the centralization index exhibits a constant increase from 2002 to 2008, suggesting a growing domination of the top destination. This trend differs from the previous study that examined the international tourist flow network (Seok et al., 2021), which may be explained by the nature of the data. Secondly, the number of clusters expands from 8 to 11 during the same period, indicating a diversification of clusters over time. As for the influential factors, the QAP regression results show that distance proximity has a positive impact. Due to the limited travel range of package tours, where buses are the primary means of transportation, this trend occurs. However, attractiveness proximity is replaced by city hierarchical proximity since 2006, leading to a negative effect on the flow. The QAP results also suggest that package tours tend to flow between distant destinations and destinations of the same level. This study provides useful insights for developing data-driven tourism policymaking.
Reference
Seok, H., Barnett, G. A., & Nam, Y. (2021). A social network analysis of international tourism flow. Quality & quantity, 55(2), 419-439.
Several studies have explored geographical networks from a spatio-temporal perspective, including transportation, urban systems, and goods supply networks. However, the evolution of tourism destination networks has received less attention, particularly on a regional or national scale due to a lack of time-series flow data. The tourism destination network is dynamic and continuously changing, and understanding its evolution is crucial in addressing a range of challenges, such as overtourism and regional revitalization. This research aims to use network analysis to estimate the spatio-temporal evolution of the tourism destination network in Japan. The study focuses on inbound tourism in the 2000s, specifically during the target period of analysis, as inbound tourism during that period had a significant impact due to the implementation of the Visit Japan Campaign policy. The research investigates three main questions:
RQ1: Will the tourism destination network become more centralized or decentralized?
RQ2: Will more clusters emerge over time?
RQ3: What are the influential factors shaping the network?
Data and Methodology:
The study considers the itinerary of package tours to construct the networks between destinations. The target market is Taiwan, which is one of the major inbound markets to Japan. The data was obtained from the "Wayback Machine," a long-term website archive. Historical package tours sold by a major travel agent during the period of 2002-2008 were collected. After manual data cleaning, an adjacency matrix of 212 city-level destinations in Japan was created.
This study adopts network analysis, a quantitative methodology rooted in graph theory that has been utilized in geography since the 1960s, to examine the relationships between actors. In this study, the destinations are treated as actors, and the touring routes are regarded as links. The analysis consists of three parts: the centralization index, which reflects the global character of the network; community detection, which reveals the local character of the network; and Quadratic Assignment Procedure (QAP), which identifies the influential factors that shape the network.
Results and discussions
The study reveals three main findings. Firstly, the centralization index exhibits a constant increase from 2002 to 2008, suggesting a growing domination of the top destination. This trend differs from the previous study that examined the international tourist flow network (Seok et al., 2021), which may be explained by the nature of the data. Secondly, the number of clusters expands from 8 to 11 during the same period, indicating a diversification of clusters over time. As for the influential factors, the QAP regression results show that distance proximity has a positive impact. Due to the limited travel range of package tours, where buses are the primary means of transportation, this trend occurs. However, attractiveness proximity is replaced by city hierarchical proximity since 2006, leading to a negative effect on the flow. The QAP results also suggest that package tours tend to flow between distant destinations and destinations of the same level. This study provides useful insights for developing data-driven tourism policymaking.
Reference
Seok, H., Barnett, G. A., & Nam, Y. (2021). A social network analysis of international tourism flow. Quality & quantity, 55(2), 419-439.