Increasing the Geographical Granularity of Economic Indicators with Google Trends [Aumentare la Granularita Geografica degli Indicatori `Economici con Google Trends]

This paper proposes a method to compute a regional Economic Sentiment Index (ESI) using Google Trends (GT) data. The ESI is a leading indicator of macroeconomic magnitudes, and GT offers a high-frequency and real-time measure of information demand. The proposed method consists of retrieving the search volumes for terms potentially correlated with the ESI, selecting those highly correlated, conducting a principal component analysis (PCA), and estimating a linear regression model. The method was applied to the ESI in Spain, and the results showed that it is possible to estimate the ESI at a regional level using the PCA factor loadings of the national data.