Computational Methods for Economists Syllabus
Course
Econ 2020: Computational Methods for Economists
Meeting Times & Location
Monday and Wednesday, 1:30 pm - 2:50 pm
see Canvas for location
Course Site
Course Description
This course prepares students to conduct independent research by providing the necessary skills in programming, project organization, and the use of AI tools for research.
The course has two halves.
- The first half is done without AI. It covers version control, project organization, programming in R, and an introduction to Python. These are the skills you need to understand your own code and data, and to judge code and output that someone (or something) else produced.
- The second half covers AI methods for research. It starts with how large language models work and how to use them as a research tool, then turns to AI agents: tools that read your files, write and run code, and act on your project. The emphasis throughout is on checking results and keeping them reproducible. Agents make producing research output nearly free, but output is only useful once it has been checked and understood, and that work now has to be done on purpose.
By the end of the course students should be comfortable programming in the major languages used in economic research, producing their own work as a reproducible project, and using AI agents in a way they can verify.
Learning Goals
- Write clean, documented, reproducible code in R, with a working knowledge of Python
- Organize a research project with version control and a clear structure
- Understand how large language models work, and use them as a research tool
- Direct AI agents on research tasks, and verify what they produce
- Apply software tools and best practices to economic research projects
Schedule
The schedule is subject to change as the course progresses.
Click on the icons for links to the lecture slides, class feedback surveys, and assignments. Assignments marked with a day are due at 11:59 pm on that day.
| # | Date | Topic | Lecture Slides | Assignments Due | Class Feedback |
|---|---|---|---|---|---|
| 1 | 1/27 | Intro, Git & GitHub | |||
| 2 | 2/01 | Project Organization & VS Code | |||
| 3 | 2/03 | Git Branches & Pull Requests | |||
| 4 | 2/08 | Base R | |||
| 5 | 2/10 | R Data Wrangling: tidyverse |
|||
| 2/15 | No Class (Long Weekend) | ||||
| 6 | 2/17 | Guest: Data Librarian [TBD: confirm] | |||
| 7 | 2/22 | R Databases and APIs | |||
| 8 | 2/24 | R Functions & Regressions | Proposal (Fri 2/26) | ||
| 9 | 3/01 | R Data Visualization | |||
| 10 | 3/03 | R Data Wrangling: data.table & Speed |
|||
| 11 | 3/08 | Python I: Basics | |||
| 12 | 3/10 | Python II: pandas |
Checkpoint (Fri 3/12) | ||
| 13 | 3/15 | Guest: OSCAR High Performance Cluster [TBD: confirm]; Web Basics | |||
| 14 | 3/17 | How LLMs Work | |||
| 15 | 3/22 | LLMs as a Research Tool: APIs & Text Classification | APIs Local | ||
| 16 | 3/24 | From Chatbots to Agents | Website (Fri 3/26) | ||
| 3/29 | No Class (Spring Recess) | ||||
| 3/31 | No Class (Spring Recess) | ||||
| 17 | 4/05 | Directing Agents | |||
| 18 | 4/07 | Reviewing Agent Work | |||
| 19 | 4/12 | Midterm (in class) | |||
| 20 | 4/14 | Reproducible Documents with Quarto |
AI Replication (Fri 4/16) | ||
| 21 | 4/19 | Checking Results: Tests & Independent Implementations | |||
| 22 | 4/21 | Research Design with Agents | |||
| 23 | 4/26 | Final Presentations | Final Project | ||
| 24 | 4/28 | Final Presentations |
We will not use the final exam slot given by the registrar. Please use the time to study for your other finals!
Class Feedback
Each lecture will have an accompanying survey for students to fill out (see the schedule). The survey will ask about comprehension of topics covered in the lecture and will have open ended space for questions. These assignments are graded for completion only and are meant as a way to judge if any material needs to be covered again in more detail. Do not fill these out if you miss a class, as they are meant to reflect participation in the class.
Assignments
Research Proposal
Early in the semester you will propose the research project you will work on for the rest of the course. The proposal describes a research question and a data source you have already opened, and includes a first pull of the data with R, committed to your project repository. The proposal is done without AI.
Final Project Checkpoint
Before spring recess you will submit a cleaned dataset and one descriptive figure for your project, made with R or Python without AI. The checkpoint is graded for completion and counts toward the final project.
Personal Website
Every academic should have a website. You will build a personal academic website with an AI agent and host it on GitHub Pages. We will start it together in class.
Midterm
An in-class exam on the first half of the course: R, Python, Git, and project organization. It is taken without AI or internet access. Most of the exam covers R.
AI Replication
You will use an AI agent to reproduce one table or figure from a published paper’s replication package, chosen from a list I provide. You will submit the reproduced result, the agent’s session transcript, and a short log of what matched, what you checked, and what the agent got wrong.
Final Project
The final project builds on your proposal. You will carry your project forward with AI agents, write it up as a reproducible Quarto document in which every number, table, and figure is produced by code, and present it in class during the last week. You will also submit your agent session transcripts and a log of the checks you ran.
Problem Sets
Optional problem sets accompany the lectures. They are not graded, and solutions will be posted. They are the best way to practice for the midterm.
AI Policy
First half (through the 3/15 class): no AI. Do not use AI tools (e.g., ChatGPT, Claude, Google Gemini, GitHub Copilot, or AI features in your editor) for coursework in the first half, including the research proposal and the final project checkpoint. The point of the first half is to build the skills you will need to judge AI output later, and those skills only come from doing the work yourself.
Second half (from the 3/17 lecture on): AI expected. You will use AI tools, including agents, for the website, the AI replication, and the final project. You are responsible for everything you submit. Submit your agent session transcripts where an assignment asks for them, and be able to explain any code or result in your work.
Grading
Each assignment will be graded out of 100 points.
| Assignment | Weight |
|---|---|
| Class Feedback | 15% |
| Research Proposal | 10% |
| Personal Website | 5% |
| Midterm | 25% |
| AI Replication | 10% |
| Final Project | 35% |
The final project grade includes the checkpoint, the paper, and the presentation.
Letter grades will be given according to the following rubric:
| Letter Grade | Numeric Grade |
|---|---|
| A | [90, 100] |
| B | [80, 89) |
| C | [70, 79) |
| NP | [0, 69) |
Late assignments will be penalized 10 points per day late. Late final presentations will lose full points, given the need to present in class. If you contact me ahead of time explaining why you need an extension, I may be able to accommodate.
Credit Hours
You are expected to spend 180 hours on this course. You will spend approximately 32 hours in lectures, 60 hours on readings, review, and the optional problem sets, 15 hours preparing for the midterm, and 73 hours on the graded assignments (about 40 on the final project).
Accessibility and Accommodations Statement
Brown University is committed to full inclusion of all students. Please inform me early in the term if you may require accommodations or modification of any of course procedures. You may speak with me after class, during office hours, or by appointment. If you need accommodations around online learning or in classroom accommodations, please be sure to reach out to Student Accessibility Services (SAS) for their assistance (sas@brown.edu, 401-863-9588). Undergraduates in need of short-term academic advice or support can contact an academic dean in the College by emailing college@brown.edu. Graduate students may contact one of the deans in the Graduate School by emailing graduate_school@brown.edu.
Acknowledgements
Material in this course comes from many locations, but I am especially thankful to the lecturers for this course before me, Matthew DeHaven, Michael Neubauer, and Shunsuke Tsuda for sharing their materials.
The framing of the second half draws on Scott Cunningham’s talk “AI Methods for Research Workers” (2026).
I also want to acknowledge the excellent lecture slides by Grant McDermott for teaching a course on R at the University of Oregon.