Reasoning with Data

Spring 2026 — COS 424

See the canvas page of this course for the most up-to-date version.

Reasoning with Data trains students to turn messy real-world datasets and experiments into credible empirical claims, covering data wrangling and visualization, core inference and causal tools (potential outcomes, DAGs, experiments, regression), and modern frontiers (LLMs, causal ML, benchmarks, and quasi-experiments). The course puts special emphasis on developing the practical skills that students will need in data-facing roles.

Schedule

Week	Lectures and Slides	Quiz?	Precept and Homework	Project
Week 01	Tuesday: No class Thursday: Introduction	✗	✗	✗
Week 02	Tuesday: Handling Data Thursday: Visualizing Data	Mock	P1: Basic Pandas	✗
Week 03	Tuesday: Describing Data Thursday: Network Data	✔	P2: Data Visualization P1 Homework Due	Work on Milestone #1
Week 04	Tuesday: Applied ML #1 Thursday: Applied ML #2	✔	P3: Networks P2 Homework Due	Work on Milestone #1
Week 05	Tuesday: Handling Text Thursday: Using LLMs	✔	P4: Text P3 Homework Due	Work on Milestone #1
Week 06	Tuesday: Potential Outcomes Thursday: Experiments	✔	Project Ideas Presentation P4 Homework Due	Milestone #1 Due
Week 07	Tuesday: DAGs Thursday: Regression #1	✔	P5: Causal Inference P4 Homework Due	Work on Milestone #2
Week 08	—	—	—	Work on Milestone #2
Week 09	Tuesday: Regression #2 Thursday: Regression #3	✔	P6: Regression P5 Homework Due	Work on Milestone #2
Week 10	Tuesday: Quasi-Experiments Thursday: Labeling with LLMs	✔	Project Office Hours P6 Homework Due	Milestone #2 Due
Week 11	Tuesday: Recommender Systems Thursday: Causal ML	✔	Project Office Hours	Work on Milestone #3
Week 12	Tuesday: Guest Lecture Thursday: Guest Lecture	✗	Project Office Hours	Work on Milestone #3
Week 13	Tuesday: Guest Lecture Thursday: No Class	✗	✗	Work on Milestone #3
Week 14	Tuesday: Guest Lecture Thursday: Final Presentation	✗	Back-up: Potentially Project Presentations	Work on Milestone #3

Final Project Deadline: Following university policy, you must submit your final project report during the 3-hour window on December 17, from 8:30–11:30 a.m. ET, as set by the University Registrar.

Course Organization

Project — 50% of grade

An end-to-end data science project completed in groups of up to 5 students. Use of generative AI is encouraged for the project.

The project consists of three milestones:

M1 — Project Proposal (2%): Propose your project idea.
M2 — Initial Analysis (8%): Initial data wrangling and results.
M3 — Final Project (40%): Final report and presentation.

Quizzes — 25% of grade

Multiple-choice questions administered at the beginning of most Thursday lectures.

Quiz retakes are not allowed.
However, your two lowest quiz grades will be dropped.

Homework — 25% of grade

Short exercises assigned at the end of each precept. Use of generative AI is not allowed for homework.

Homework is graded on a fail/pass basis.
Pass = at least 60% of answers correct.
Released: Fridays.
Deadline: The following Friday at the end of the day, with some exceptions.
Late submissions are not allowed.
However, your lowest homework grade will be dropped.

Precepts

Precepts are not graded, but they are an essential part of the course. They are a place where you can:

Ask questions about homework and projects.
Work through tutorials that teach practical data science skills.
Practice concepts and techniques that will appear in homework and the project.

Course Policies

Diversity and Respectful Conduct

This course welcomes students of all backgrounds. You should expect and demand respectful treatment from your classmates and instructors. If any incident challenges our commitment to a supportive, diverse, inclusive, and equitable environment, please let the instructors know so that we can address the issue.

Late Submissions and Quiz Retakes

Late submissions and quiz retakes are not allowed. To accommodate occasional difficulties, however, we will:

Drop your two lowest quiz grades.
Drop your lowest homework grade.

The final project milestone submission date is set by the University Registrar, so we cannot accept late submissions for the final milestone. For emergencies affecting the other two project milestones, please consult the instructors.

Disability, Religious, and Family Accommodations

If you have questions about disability or religious accommodations, please refer to the relevant university policies. You are also welcome to contact the instructors about accommodations or other circumstances that may affect your participation in the course.

Academic Integrity

We will follow the University’s Rules and Responsibilities guide.

Collaboration with Classmates

We encourage you to discuss course material with classmates, either privately through personal interactions or publicly on Ed.

For homework, however, all solutions must be written independently. You may not copy another student’s homework assignment or allow another student to copy yours.

Use of Generative AI

The use of generative AI tools (e.g., ChatGPT, Gemini in Colab, GitHub Copilot) depends on the course activity:

Course project: Generative AI use is allowed and encouraged.
Precepts and homework: Generative AI use is not allowed.

The precepts and homework are designed to give you direct practice with the core skills taught in the course. Much of the supporting code is already provided, so the amount of coding required from you is intentionally limited. Using generative AI to complete these exercises would undermine their learning objectives.

Acknowledgements

This course borrows heavily from EPFL’s Applied Data Analysis course. Thank you, Bob!