+49 (0)6221 54-2958
joep.lustenhouwer@awi.uni-heidelberg.de

University of Heidelberg
Alfred-Weber-Institute of Economics
Bergheimer Str. 58
D-69115 Heidelberg
Germany

Mini-workshop on natural language processing (NLP) for economists

On Wednesday November 9th and Thursday November 10th 2022, Julian Ashwin ( London Business School ) will give a workshop on natural language processing for economists. The workshop takes place at Heidelberg University and consists of 4 Lectures. Participation from external PhD students and postdocs is possible but limited. To participate, please register by sending an email to joep.lustenhouwer@awi.uni-heidelberg.de.

Course description

Over four lectures, we will explore several approaches to natural language processing (NLP) for economists, including practical steps on implementation. In the first lecture, we will cover dictionary-based measures with a particular focus sentiment analysis for forecasting economic growth. In the second lecture, we will cover topic modelling and other unsupervised methods, motivated by an application to central bank communication. The third lecture will focus on supervised NLP methods, illustrated by an example scaling up qualitative analysis of open-ended interviews with Rohingya refugees. Finally in the fourth lecture we will cover approaches to NLP for causal analysis with particular attention on Bayesian Topic Regression and an application to the effect of media coverage on financial volatility. 

Schedule

Wednesday Nov. 9

10:00-11:30 hours –  Lecture 1: Dictionary methods and forecasting economic growth with news sentiment

Thursday Nov. 10

9:00-10:30 hours – Lecture 2: Topic modelling and central bank communication

10:45-12:15 hours –  Lecture 3: Supervised NLP and qualitative analysis of unstructured interviews

14:00-15:30 hours – Lecture 4: Causal NLP and the financial volatility effects of media coverage