Big Data in Macroeconomic Analysis: Texts as data source for the economic analysis - Potentials and possible applications

Current Project

Project Management

Dr. Claus Michelsen

Project Period

August 15, 2019 - March 15, 2021

Commissioned by

German Federal Ministry for Economic Affairs and Energy

In Cooperation With

JULIE Lab, Friedrich-Schiller University of Jena

The aim of this project is to use innovative approaches and heterogeneous data sources to advance research on forecasting economic development, predicting turning points and determining economic cycles. The first work package provides a comprehensive overview in the field of automated text analysis, the use of media data in the economic forecast and the methods of forecast evaluation, as well as the possible data sources. The second work package builds on these findings and applies promising analytical approaches to German-language data sources. The first focus is on the analysis of information from newspaper articles. It examines and discusses a transfer of these approaches to information from social media, such as Twitter, Facebook or Linkedin. The indicators obtained are used in the third work package to determine and forecast economic development, its turning points and economic cycles. It will be examined whether and in what context these data can lead to an improvement of the forecasting models, which lead time can these provide in the forecast of the economic development and if they possibly provide early indications for economic recoveries respectively recessions than conventional indicators.

DIW Team

Contact

Tatjana Ribakoff

Research Project Coordinator in the Macroeconomics Department