Nowcasting with online search queries


The famous study by Ginsberg et al. (2008) about the use of Google trends to nowcast outbreaks of influenza demonstrated the informative value of online search behavior for many phenomena. The term nowcasting, as opposed to forecasting, emphasises the aim to infer information about the current state or performance of events. Nowcasting is applied to many areas where time is pressing, but publications through other means, like official statistics, usually come with a delay. Main idea is that online search or blogging activity of people is informative about commercial transactions, doctor visits, use of services, or unemployment. Over the past few years, several studies have been realized addressing the development of consumer confidence in Italy and the Netherlands (Giovanni Fassino), retail sales (Micha Vink), unemployment throughout Europe (Maaike Gerritse), Dutch import and exports (Thomas Coebergh van den Braak), Dutch housing sales (Jan van Werkhoven) and quarterly sales of S&P100 firms (Matthijs Versteeg). Apart from taking care of the forecasting, these studies are concerned with the collection of online data and the selection of useful predictors. Particularly challenging in this research is the identification of relevant information in the enormous amount of data available each time instant. 

External Partners

--

Timing

This project is ongoing.

Working/Published Papers

Coebergh van den Braak, Th., 2014, Nowcasting export and import figures in the Netherlands, MSc thesis, Rotterdam School of Management, Erasmus University Rotterdam.

Fassino, G., 2013, Nowcasting consumer confidence with Google trends: the case of Italy and the Netherlands, MSc thesis, Rotterdam School of Management, Erasmus University Rotterdam.

Gerritse, M., 2014, Nowcasting: how use Google trends to predict unemployment in Europe, MSc thesis, Rotterdam School of Management, Erasmus University Rotterdam, MSc thesis, Rotterdam School of Management, Erasmus University Rotterdam.

Versteeg, M., 2015, Forecasting quarterly sales with Google Trends, MSc thesis, Rotterdam School of Management, Erasmus University Rotterdam.

Vink, M., 2014, Using Google trends to nowcast retail sales in the Netherlands, MSc thesis, Rotterdam School of Management, Erasmus University Rotterdam.

Werkhoven, J. van, 2014, Nowcasting Dutch housing sales using Google trends, MSc thesis, Rotterdam School of Management, Erasmus University Rotterdam.

PhD Thesis

--

Video/Interview

--

For more information please contact:

Jan van Dalen

Director

Tel: + 31 (0)10-408 2370
E-mail: jdalen@rsm.nl