Forecasting Tourism Demand with Google Trends For a Major European City Destination

Abstract

The purpose of this study is to investigate whether using Google Trends indices for web and image search improves tourism demand forecast accuracy relative to a purely autoregressive baseline model. To this end, Vienna—one of the top-10 European city destinations—is chosen as a case example for which the predictive power of Google Trends is evaluated at the total demand and at the source market levels. The effect of the search query language on predictability of arrivals is considered, and differences between seasonal and seasonally adjusted data are investigated. The results confirm that the forecast accuracy is improved when Google Trends data are included across source markets and forecast horizons for seasonal and seasonally adjusted data, leaning toward native language searches. This outperformance not only holds relative to purely autoregressive baseline specifications but also relative to time-series models such as Holt–Winters and naive benchmarks, in which the latter are significantly outperformed on a regular basis.