Search

Tourism involves the movement of people. Visitors move between destinations differently, which creates patterns and networks that link destinations within travel itineraries. These patterns contain valuable information that can be analysed for tourism product development, to improve marketing services and assist in infrastructure and transportation development decisions. Traditionally, research on visitors' travel patterns relied on data collected from interviews, surveys, or direct observation. Normally the size of this type of data is relatively small, and processing of data collection is costly as well as time consuming. The introduction of big data analytics within tourism research has seen a diversification within data sources, and social media has become a particularly valuable source. Social media generates vast volumes of data, which in addition to text, images, and video content contains additional information in the form of metadata, such as geo-location (where posts were made) and user origins. When combined with advances in big data analytics, this type of data opens an opportunity to model temporal and spatial travel patterns and especially understand the travel pattern drivers behind the travel patterns. Due to the flexibility of social media data collection, understanding visitors' travel patterns is not limited to one destination but can be expanded to the national or global scale. This thesis took advantage of the value of social media data and big data analytics, in five independent but interconnected chapters to report, and advance knowledge on visitors travel patterns. The thesis is partly comprised of papers with a compilation of five manuscripts in addition to an introduction, an overview of the methodology, and a conclusion chapter. Manuscript one consisted of a literature review that synthesises the relevant literature relating to social media, tourism, and travel patterns. In the review, key parameters were defined to describe and synthesise the emerging field of tracking visitor flows with social media data. The review discussed commonly used methods and technologies and makes recommendations for future research approaches. Manuscript two focused on data analytics and theory validation. It applied social network analysis and core-periphery theory using social media data to build the travel networks. The case study followed the travel of Chinese visitors between selected Australian destinations to examine travel structures. The Chinese social media platform Weibo was the data source. The results demonstrated that Chinese visitors in Australia mostly visit iconic Australian destinations first, before visiting destinations classified as semi-core or periphery. The findings indicated the suitability of applying core-periphery theory to social media data. It was also revealed that consideration should be given to the number of destinations within visitors' travel itineraries when analysing the core-periphery travel structures. The rationale for this focus was that the same destination may play a different role if the number of destinations is different. Manuscript three explored global tourist mobility from social media data and expanded the focus study to a global perspective. The case study investigated travel patterns of Chinese visitors before and after travel to Australia. This work addressed global tourist mobility and how Australia is positioned as an international travel destination. Results showed that Chinese visitors visited thirteen regions before or after their travel to Australia, and they appeared to travel directly from Australia to China, which indicated that Australia was the most prominent gateway country. The proposed method for identifying global travel patterns may be helpful to understand the impacts of the coronavirus pandemic and highlights that tracking mobility is important to understand the spread of diseases as well as opportunities for tourism recovery. Manuscript four proposed a method that identified Chinese travel sentiment in Chinese language. The method was tested using data from Weibo, and a case that focused on Weibo users’ posts related to visiting the Great Barrier Reef (GBR) was conducted. The manuscript detailed the process of capturing the Weibo posts describing the creation of lexicon and presented an algorithm for sentiment calculation. By investigating the sentiment towards different GBR destinations, it was demonstrated that the proposed method was effective in obtaining insights and may be suitable to monitor visitor opinions. Manuscript five introduced a new methodology that conducted importance-performance analysis (IPA) to understand visitor satisfaction with travel patterns. The new method indirectly measured performance and importance values from publicly available social media data (Weibo). A lexicon-based method for sentiment analysis was applied to identify performance value. Importance was calculated as the adjusted association rule mining algorithm Support, which assessed the frequency of posts at each destination. The results indicated that the proposed methodology offers an opportunity to conduct IPA indirectly from social media data with larger sample sizes, but at a lower cost with flexible data collection. The research used Chinese visitors to Australia as a case study, but the proposed approach could benefit a range of decision-makers directly or indirectly related to tourism.