Abstract
This study examines the influence of
online search behavior—specifically keywords related to recreational activities
popular among Taiwanese domestic travelers—on total tourism expenditure.
Utilizing high-frequency data from Google Trends, the analysis investigates how
public interest in specific leisure activities correlates with tourism
spending. Principal Component Analysis (PCA) is used to extract key indicators
representing aggregated search trends, which are then incorporated into a
Vector Autoregression (VAR) model to assess their dynamic relationship with
tourism expenditure. Results indicate that among four major categories—gourmet,
nature sightseeing, other leisure, and cultural experiences—six keywords
(“drinking coffee,” “whale watching,” “shopping,” “farm,” “indigenous culture,”
and “boating”) significantly affect domestic tourism expenditure. These
activities contribute to broader consumption in areas such as lodging,
transportation, and dining. The study contributes by (1) demonstrating the predictive
utility of online search data for tourism economics, and (2) highlighting the
growing significance of specific recreational activities, consistent with
official statistics.
JEL classification numbers: C60, O11, R11.
Keywords: Principal Component
Analysis (PCA), Recreational Activities, Tourism Expenditure, Vector
Autoregression (VAR).