In this paper, we examined the relationship
between tourism and service consumption in Taiwan. The service consumption in
Taiwan is nowcasted with the real-time tourism data in Google Trends database.
We used the high-frequency internet-searching tourism data to predict the
low-frequency service consumption data, for the real-time data with rich
information could enhance prediction accuracy. Applying the Principal
Components Analysis (PCA), we used the internet-searching tourism keywords in
Google Trends database to construct the diffusion indices. Following the
classification of the tourism keywords in Matsumoto et al. (2013), we
classified those keywords into five groups and twenty-nine classifications. We
focused on the reciprocal reactions between those diffusion indices with
service consumption to conclude which component has higher influence on service
consumption in Taiwan. Our empirical results indicated that the keywords in
“Recreational areas, and Travel-related” group have significant effects on
service consumption in Taiwan via nowcasting. Among the components of those
diffusion indices, “Farm, Travel insurance, and Visitor center” are important
variables with higher weights in common.
JEL classification numbers: C60, C80, E01, E2, E60.
Keywords: Nowcasting, the
Principal Components Analysis (PCA), Service Consumption, Tourism.