ECO 395M: Data Mining and Statistical Learning

Welcome to the Spring 2024 edition of ECO 395M, a course on data mining and statistical learning for students in the Master's program in Economics at UT-Austin. All course materials can be found through this GitHub page. Please see the course syllabus for details about:

• expectations
• assignments and grading
• readings
• other important administrative information

The exercises will be posted here as they are assigned throughout the semester.

Office hours

Mondays and Wednesdays 1-2 PM in Welch 5.228G. The best way to get here is to enter into Welch Hall through the glass doors along the Speedway side of the building. You should be in a large atrium with a large video screen on the far wall. Please use the elevator on the rightmost (northern) side of this atrium and head up to the 5th floor. Take a left out of the elevator and then an immediate right down the hallway. Enter through the main office in Welch 5.216, which has big-frosted glass walls. Proceed from there to the hallway with my office.

Topics outline

I assume that you start the semester with a basic understanding of R and data visualization, at the level of Lessons 1-5 of Data Science in R: A Gentle Introduction. This material was covered in ECO 394D, and although we'll review some of these skills in the course of learning new stuff, it's expected that you're familiar with these lessons from day 1.

The data scientist's toolbox

Slides here.

Topics: Good data-curation and data-analysis practices; R; Markdown and RMarkdown; Jupyter; the importance of replicable analyses; version control with Git and Github.

Resources to learn Github and RMarkdown:

• Introduction to RMarkdown and RMarkdown tutorial
• Introduction to GitHub
• Getting starting with GitHub Desktop

Basic concepts in statistical learning

Slides here.

Reading: Chapters 1-2 of "Introduction to Statistical Learning."

In class:

• loadhou.R
• loadhou.csv
• utilities.R
• utilities.csv

Linear models

Slides here.

Reading: Chapter 3 of "Introduction to Statistical Learning."

In class:

• oj.R and oj.csv
• saratoga_lm.R
• feature_engineer.R)

Classification

Slides here.

Reading: Chapter 4 of "Introduction to Statistical Learning."

In class:

• spamtoy.R
• spamfit.csv and spamtest.csv
• glass.R
• glass_mlr.R
• congress109_bayes.R
• congress109.csv
• congress109members.csv
• glass_LDA.R

Model selection and regularization

Slides here.

Reading: chapter 6 of Introduction to Statistical Learning.

In-class:

• saratoga_step.R
• semiconductor.R and semiconductor.csv
• hockey.R and all the files in data/hockey/
• gasoline.R and gasoline.csv

Trees

Slides here.

Reading: Chapter 8 of Introduction to Statistical Learning.

The pdp package for partial dependence plots from nonparametric regression models.

Clustering

Slides here.
Reading: chapter 10.3 of Introduction to Statistical Learning.

In class:

• cars.R
• hclust_examples.R
• linkage_minmax.R

Dimensionality reduction: PCA and tSNE

Principal component analysis (PCA). T-distributed stochastic neighbor embedding (tSNE).

Reading: rest of chapter 10 of Introduction to Statistical Learning.

Slides on PCA here.

• pca_intro.R
• nbc.R, nbc_showdetails.csv, nbc_pilotsurvey.csv
• congress109.R
• ercot_PCA.R, ercot.zip
• tSNE.ipynb

Neural networks: the basics

Intro to neural network slides here. Jupyter notebooks here.

Unsupervised learning: networks and association rules

Intro slides on networks.

Further slides on networks.

Slides on association rules here.

Miscellaneous:

• Gephi, a great piece of software for exploring graphs
• The Gephi quick-start tutorial

Scripts and data:

• microfi_gamlr.R
• microfi_households.csv
• microfi_edges.txt

Treatment effects

Treatment effects; multi-armed bandits and Thompson sampling; high-dimensional treatment effects with the lasso.

Slides:

• Treatments.

Scripts and data:

• mab.R and Ads_CTR_Optimisation.csv
• abortion.R and abortion.dat
• smallbeer.R and smallbeer.csv

Text

Slides on text.

• tm_examples.R
• smallbeer.R