Weekly Materials & Readings

Data Science and Public Policy, McCourt School for Public Polcy, Georgetown University

Week 1, September 2nd

• Topics: Course overview, logistics, introduction to R

• Lecture: Week 1 - Introduction

• Coding Materials: Week 1 - Introduction to R

• Data: Pooling Data

Week 2, September 8th

• Topics: How to represent text as data, Bag of Words Assumption, Tokens and Pre-Processing

• Lecture: Week 2 - From Text to Matrices: Representing Text as Data

• Coding Materials: Week 2 - Text Data Processing

• Data:

  • Tweets from Elections

Week 3, September 15

• Topics: vector space model, comparing documents, textual complexity.

• Lecture: Week 3 - Descriptive Inference

• Coding Materials: Week 3 - Descriptive Inference

Week 4: September 22

• Topics: Dictionaries, Pros and Cons, and Using Off-the-Shelf Models

• Lecture: Week 4 - Dictionaries and Off-the-Shelf Classifiers

• Coding Materials:

  • Week 4 - Dictionaries and Off-the-Shelf Classifiers

  • Week 4 - Transformers Based Classifiers

  • Week 4 - Transformers Based Classifiers in R , html here

• Data: Streamming Chat Comments

Week 5: September 29

• Topics: Building a training data, regularization, model selection, and model evaluation.

• Lecture: Week 5 - Supervised Learning

• Coding Materials: Week 5 - Supervised Learning: Training your own classifiers

Week 6: October 6

• Topics: Unsupervised learning, topic models, and clustering

• Lecture: Week 6 - Unsupervised Learning

• Coding Materials: Week 7 - Unsupervised Learning: Topic Models

Week 7: October 20

• Topics: Unsupervised learning + Scaling

• Lecture: Week 7 - Scaling

• Coding Materials: Week 10 - Scaling

Week 8: October 27

• Topics: Introduction to Deep Learning

• Lecture: Week 8 - Deep Learning

• Coding Materials (OPTIONAL):

  • Logistic Regression as a Neural Network

  • Introduction to Pytorch

Week 9: November 3

• Topics: Word Embeddings - Theory

• Lecture: Week 9 - Word embeddings

• Coding Materials in R:

  • Week 9 - Word Embeddings in R

• Coding Materials in Python (Optional)

  • Word Vectors with Gensim: notebook

Week 10: November 10

• Topics: Word Embeddings - Applied Paper

• Lecture: Week 10 - Word embeddings II

• Coding Materials in R:

  • Week 9 - Word Embeddings in R

• Coding Materials in Python (Optional)

  • Word Vectors with Gensim: notebook

Week 11: November 17

• Topics: Replication Class

• Materials: Check the paper your colleagues are replication here

Week 12: November 22

• Topics: Transformers

• Lecture: Week 12 - Transformers

• Coding: Below you have more code than we will cover in class. It is a mix between code I wrote, and code from my colleague (Sebastian Vallejo) from a class we co-taught in the summer

  • Off-the-shelf transformers

  • Fine-tunning transformers

  • Further Training Transformers

  • LLM as annotators