Accurate forecasting of tourism demand is of utmost relevance for the success of
tourism businesses. This paper presents a novel approach that extends autoregressive
forecasting models by considering travellers’ web search behaviour as additional
input for predicting tourist arrivals. More precisely, the study presents a method
with the capacity to identify relevant search terms and time lags (i.e. time difference
between web search activities and tourist arrivals), and to aggregate these time
series into an overall web search index with maximal forecasting power on tourism
arrivals. The proposed approach enables a thorough analysis of temporal relationships
between search terms and tourist arrivals, thus, identifying patterns that reflect
online planning behaviour of travellers before visiting a destination. The study is
conducted at the leading Swedish mountain destination, Åre, using arrival data and
Google web search data for the period 2005–2012. Findings demonstrate the ability
of the proposed approach to outperform traditional autoregressive approaches, by
increasing the predictive power in forecasting tourism demand.
Springer, 2019. Vol. 21, no 1, p. 45-62
Google Trends data, Search word analysis, Online search pattern, Tourist arrival prediction, Autoregressive time series forecasting, Big data