Predicting the Present Revisited: The Case of Thailand
Abstract
Google is currently the most-used search engine in the world. There are approximately 3.5 billion searches being conducted on Google each day. With real-time processing, Google Trends data can be used in a prediction technique called nowcasting (or “predicting the present”) – using the current period's real-time information to estimate the current period's indicators of interest. In this paper, we showed how Google Trends can be used for nowcasting Thailand's various economic indicators. The sectors being analyzed are (i) the labor market sector (unemployment rate and unemployment registration), (ii) the real sector (automobile sales), and (iii) the financial sector (SET index). The results revealed that incorporating the Google Trends data into the prediction models improved the Adjusted R-Squared and improved the predication accuracies under various measures.
