Syllabus and Assessment

Course meets 6 times: 5th, 6th, 7th October; 11th, 12th, 13th October; 3-6 pm CEST.

Mode of meeting: Zoom (Link sent by email to registered students)

The course only has Pass/Fail grade.

Pass: attending and participating in the classes, presenting a brief discussion of one of the papers from the reading list, writing a short term paper (2-3 pages) discussing how NLP could be useful for economics research questions of your interest.

Topics:

(Each topic is one lecture, 3 hours duration)

1. Introduction

  • Course overview
  • Introduction to NLP
  • NLP, Machine Learning, and Economics: an overview

Readings:
(Note that you are not obligated to read everything thoroughly).

  • Chapter 1 from “Speech and Language Processing” by Jurafsky and Martin (available online)
  • Gentzkow, M., Kelly, B., \& Taddy, M. (2019). Text as data. Journal of Economic Literature, 57(3), 535-74.

Form: Lecture + discussion

2. Python fundamentals

  • Overview of Python
  • Basic text analysis with Python

Readings:

  • “Python for Everybody” by Charles Severence. \url{https://www.py4e.com/html3/}. The content covered in this Chapter is taken from the first 10 chapters in the book.
  • Chapter 1-2 in the NLTK book

Form: Hands-on exercises + discussion

3. NLP and Machine Learning methods

  • Overview of NLP methods
  • Examples of using existing NLP tools
  • How do NLP methods work with Economics research?

Form: Lecture + Hands-on exercises

Readings:

4. Diving deeper: Text Classification and Topic Modeling

  • Overview of text classification
  • Overview of topic modeling
  • Code walkthroughs

Form: Lecture + Hands-on exercises
Readings: Chapter 4 in Practical Natural Language Processing

5. NLP without annotated data

  • Overview of methods that don’t rely on large training datasets
  • A case study with different ways of solving one such problem
  • Code walkthroughs Form: Lecture + Hands-on exercises

6. NLP and Economics: selected readings

Form: Student presentations (perhaps in groups of 1-3 people)

You can choose from some of these papers

Expectation: Give a short overview of

  • what the paper is about?
  • what aspect of economics does the paper deal with?
  • how does it use NLP for that purpose?
  • what do you think of the paper? what are some potential limitations to the approach? etc.

Duration: 10 minutes + 5 min discussion (Not a strict requirement)

7. Student term papers

Briefly summarize (2-3 pages) what you learnt about the intersection of NLP and Economics by taking this course, and note down some thoughts on how it is useful for your own research topics. Depending on the time and interest, we can decide whether we want to have a presentation session or just writeup submissions.

Deadline: 27th October 2022

7. Recap (optional session on 17th or 18th October)

  • Discussion on topics covered
  • Review of exercises
  • Resources for the future
  • Anything you want to ask