Tourism demand forecasting with online news data mining. (September 2021)

• Record Type: Journal Article
• Authors: Park, Eunhye; Park, Jinah; Hu, Mingming
• Is Part Of: Annals of tourism research. Volume 90(2021)
• Publication Date: 2021-09
• Subjects: News discourse -- Topic modeling -- Tourism demand forecasting -- Hong Kong; Tourism -- Periodicals
• DOI: 10.1016/j.annals.2021.103273
• Languages: English
• ISSNs: 0160-7383

Abstract

Abstract: This study empirically tests the role of news discourse in forecasting tourist arrivals by examining Hong Kong. It employs structural topic modeling to identify key topics and their meanings related to tourism demand. The impact of the extracted news topics on tourist arrivals is then examined to forecast tourism demand using the seasonal autoregressive integrated moving average with the selected news topic variables method. This study confirms that including news data significantly improves forecasting performance. Our forecasting model using news topics also outperformed the others when the destination was experiencing social unrest at the local level. These findings contribute to tourism demand forecasting research by incorporating discourse analysis and can help tourism destinations address various externalities related to news media. Highlights: The study forecasts tourism demand to Hong Kong from Mainland China and US. Key topics of news discourse are extracted and employed to predict tourism demand. The structural topic model is applied to discover key topics in major newspapers. Forecasting model with news topics outperforms for normal and crisis periods. The study confirms the value of local news coverages in tourism demand forecasting.