The queries to Google Search as predictors of migration flows from Latin America to Spain
Abstract
Description
Recently, the development of global network and ITC technology provided
new opportunities to improve the estimations and predictability of migration flows.
The activity of users of e-mail and other web-based services was compared in time
and space in order to track international human mobility. At the same time, the IP
based geolocation linked to Google Search proved to be efficient in geographically
tracking the outbreaks of several illnesses, and also in predicting changes in economic
indicators and travel patterns. This research draws from both experiences. It compares
the popularity of migration-to-Spain related queries introduced to Google Search in
Argentina, Colombia and Peru, to changes in a quantity of residents’ registrations in
Spain, performed by immigrants proceeding from these countries between the years
2005 and 2010. Following the preliminary visual trend analysis, the time series are
pre-whitened in order to formally test for a time-shifted correlation and predictability
not-influenced by a general series trend. The analysis was performed on the datasets
of queries popularity derived from Google Trends and anonymized micro-data of
Residential Variation Statistics based on the Municipal Register of Spain. The
predicted lags of one or more months that showed to be significantly correlated
according to the Cross-Correlation Function have been further used to evaluate its
predictability with regression analysis. The results show a significant correlation and
weak to moderate predictability for the lags of several months depending on the
particular country. The findings support the assumption that popularity of queries to
Google Search provided by Google Trends might constitute a useful predictor of
migration flows while at the same time it indicates further developments necessary in
order to improve its analytical capacities.
PDF; 109 pgs.
PDF; 109 pgs.