Oxana Babecká Kucharčuková, Jan Brůha, Petr Štěrba
This paper introduces a novel travel performance indicator derived from tourist reviews available online, utilizing text mining techniques. The time series generated is integrated as an explanatory variable into a small-scale empirical model of travel revenue and expenditure in the Czech Republic’s balance of payments. The significance of online reviews for nowcasting is validated through various machine learning algorithms. The study also addresses empirical challenges, including trends in review data, the impact of the COVID-19 pandemic, and occasional methodological changes in official statistical series, and outlines strategies to overcome these obstacles. The findings suggest that the proposed model is a valuable addition to the Czech National Bank’s nowcasting framework. To the best of our knowledge, this is the first study to combine text analysis with nowcasting of a BoP item, specifically travel services.
JEL kódy: C53, C83, F17
Klíčová slova: Balance of payments, text mining, travel services
Vydáno: listopad 2025
Ke stažení: CNB WP No. 13/2025 (pdf, 1,6 MB)