Forecasting migration of EU citizens to Germany using Google Trends

The study examines the potential of Google Trends data as an additional data source for forecasting EU migration to Germany.For that aim, candidate search queries with relation to migration intent are proposed.The resulting Google Trends Indices (GTI) are combined with macroeconomic and past migration data and used to build a machine learning regression model.It is shown that GTI predictors can moderately reduce the forecast error and enable a slight expansion of the forecast horizon.However, the presence of outliers emphasizes the need for continuous improvement in data quality to increase the robustness of the approach.