Spurious patterns in Google Trends data - An analysis of the effects on tourism demand forecasting in Germany.

Previous studies show that time series data about the frequency of hits for tourism-related search terms from Google (Google Trends data) is a valuable predictor for short-term tourism demand forecasting in many different tourism regions worldwide. The paper contributes to this literature in three ways. First, it shows that Google Trends data is useful for short-term predictions of monthly tourist arrivals in several German holiday regions. Second, the paper also demonstrates that the Google Trends time series we employ share certain patterns with Google Trends time series used in previous studies, including several studies totally unrelated to the tourism industry. We refer to these artefacts as "spurious patterns" and perform a detailed analysis of their negative impact on forecasting. Last, the paper proposes a method to sanitize Google Trends data and reduce the adverse impact of spurious patterns, thereby paving the way to develop statistically sound tourism demand forecasts. • Google Trends data is a valuable predictor for tourism in Germany. • The data shows "spurious patterns" that are unrelated to tourism. • The spurious patterns have a negative impact on tourism forecasts. • Data modi_cation can reduce the impact of the spurious patterns. • The spurious patterns can be seen in several previously published studies. [ABSTRACT FROM AUTHOR]