36-613

Data Visualization

Fall 2026

Instructor: Shannon Gallagher (sgallagh@stat.cmu.edu)
Office: 129D Baker Hall
Office Hours: Mondays, 11:00 a.m.–noon, and Tuesdays, 1:00–2:00 p.m., in Baker Hall 129D

Teaching Assistant: Michele
TA Office Hours (Michele): Tuesdays, 2:00–3:00 p.m., in Wean Hall 3711

Lectures: Mondays and Wednesdays, 9:30–10:50 a.m., Baker Hall 140A

Course Website: https://stat.cmu.edu/~sgallagh/courses/mads-fall-2026/datavis.html
Canvas: Official announcements, summary grades, recordings, checkpoints, and restricted materials
Gradescope: Homework and final-project submission through assignment links in Canvas, detailed feedback, and regrade requests
Piazza: Course questions and visualization critiques; access through Canvas
Course Communication: Official announcements are posted on Canvas; course and assignment questions belong on Piazza
Primary Tools: R, RStudio, Quarto/R Markdown, tidyverse, ggplot2
Teaching Dataset: NOAA Storm Events Database
Final-Project Dataset: Global data, AI, and machine-learning salary archive, 2020-2025

Opening week. Monday, August 24 meets remotely at the scheduled time. Wednesday, August 26 is asynchronous; a recorded lecture will be posted on Canvas.

Two-dataset structure. NOAA Storm Events remains the shared dataset for lectures and checkpoints. Professor McGovern's final project is Global Data & AI Workforce Compensation Analytics, using a separate salary archive to support an executive-search recruiting recommendation.

Textbooks and References

No textbook is required. The following references may be useful throughout the course:

Course Description

Graphical displays of data help us understand evidence and communicate claims. In this mini-course, students will learn how to create, critique, and explain statistical graphics for real data, with an emphasis on client-facing communication. The central course dataset is NOAA Storm Events, a public record of severe weather events, including event type, time, location, injuries, fatalities, damage estimates, physical magnitude, and narrative descriptions.

The course uses a shared NOAA teaching sequence. Students will work with categorical, quantitative, temporal, spatial, and textual variables; create graphics using R and ggplot2; and connect visual evidence to statistical summaries or formal analyses where appropriate.

The course culminates in a client-style workforce-compensation report. Students advise an executive-search steering committee, select a defensible target employer, identify priority data-and-AI talent lanes, and communicate what the supplied salary archive can and cannot support. The goal is not only to make attractive plots, but to produce graphics that support decisions, reveal tradeoffs, and communicate honestly with a non-technical audience.

Course Objectives

By the end of the course, students should be able to:

  1. Understand data and reproducible analysis.

    • Distinguish among categorical, ordinal, quantitative, temporal, spatial, and text variables.
    • Choose graphics and summaries appropriate for each variable type.
    • Use R, RStudio, and Quarto/R Markdown to create reproducible analyses.
    • Document data sources, cleaning decisions, and limitations.

  2. Create statistical graphics.

    • Build readable graphics with ggplot2.
    • Create categorical displays, distribution displays, maps, time series, and selected advanced or alternative visualizations.
    • Incorporate uncertainty, statistical tests, model summaries, or other evidence when those additions strengthen the visual argument.

  3. Communicate with technical and non-technical audiences.

    • Write clear captions and interpretations.
    • Explain what a graph shows, what it does not show, and why it matters.
    • Present visual findings as part of a coherent client recommendation.

  4. Critique statistical graphics.

    • Identify strengths and weaknesses of graphical choices.
    • Give constructive feedback on peers' visualizations.
    • Revise graphics based on critique, audience, and purpose.

Tentative Course Timeline

Week Topic Major Work
Week 0 Setup, R/RStudio/Quarto, NOAA starter data Starter file
Week 1 Visualization principles, tidy data, grammar of graphics, categorical displays HW1 opens Sun Aug 23
Week 2 Categorical comparison; quantitative distributions, ECDFs, density HW2 opens Sun Aug 30; HW1 due Tue Sep 1
Week 3 Scatterplots, trends, residuals, contours, and binned displays HW3 opens Sun Sep 6; HW2 due Tue Sep 8; Critique 1 due Fri Sep 11
Week 4 Multivariate displays, distance, MDS, and PCA HW4 opens Sun Sep 13; HW3 due Tue Sep 15
Week 5 Nonlinear embeddings, time-series structure, and spatial foundations HW4 due Tue Sep 22
Week 6 Areal maps, graphic design, and text In-class checkpoints
Week 7 Animation, interaction, critique, and client communication Critique 2 due Wed Oct 7; final HTML report due Fri Oct 9

The final project is due Friday, October 9 at 11:59 p.m. Eastern.

Course Components

1. Lectures

Lectures introduce the main visualization concepts and show worked examples in R. Students are expected to attend, take notes, ask questions, and participate in discussion. Unannotated slides, example code, and supporting materials are posted on the course website. Annotated slides are released through Canvas only after the homework associated with that material is due. Recordings are also provided through Canvas. If you miss class, you are responsible for the material covered that day.

2. Checkpoints

The course includes occasional short coding, interpretation, or visualization checkpoints. These activities use NOAA Storm Events, prepare students for the homework and final project, and together count for 20% of the course grade. When a checkpoint requires an electronic response, it is provided and submitted through Canvas.

3. Homework

Homework assignments follow a four-part sequence. NOAA Storm Events is used primarily as the shared example in lectures and in-class checkpoints. Each assignment should be completed in Quarto/R Markdown unless otherwise specified.

Current homework sequence:

Homework files are released on the teaching website at 12:00 a.m. Eastern on Sunday and are due the following Tuesday at 11:59 p.m. Eastern:

Assignment Opens Due
HW1 Sun Aug 23 at 12:00 a.m. Tue Sep 1 at 11:59 p.m.
HW2 Sun Aug 30 at 12:00 a.m. Tue Sep 8 at 11:59 p.m.
HW3 Sun Sep 6 at 12:00 a.m. Tue Sep 15 at 11:59 p.m.
HW4 Sun Sep 13 at 12:00 a.m. Tue Sep 22 at 11:59 p.m.

Homework instructions and files are released through the teaching website. Open the Gradescope assignment from Canvas and submit each homework as a single PDF. Detailed feedback and regrade requests are handled in Gradescope; summary grades are recorded in Canvas. Solutions are released only through Canvas after the deadline and may not be publicly distributed.

Time series, maps, text, animation, interaction, and critique are assessed through the occasional checkpoints and critiques rather than additional full homework assignments.

Assignments should be written clearly, with graphs labeled and interpreted in complete sentences. Code should be readable and reproducible. Only the rendered PDF is submitted for homework unless an assignment explicitly states otherwise.

4. Graphics Critique

Students will practice critiquing visualizations found in the world: news, public reports, dashboards, advertisements, consulting reports, or scientific communication. For a critique, students should identify the source, describe what the graphic shows, explain its strengths and weaknesses, and suggest one concrete improvement.

Students complete two graphics critiques, each worth 10% of the course grade, as Piazza posts visible to the class. Each post must include or link to the visualization, identify its source, describe what it shows, explain its strengths and weaknesses, and suggest one concrete improvement. Classmates may view and discuss the posts. Do not post homework solutions, restricted data, or Canvas-only course materials.

The graph-critique problem in HW1 serves as a warm-up for the two Piazza critiques; it contributes only to HW1 and is separate from the two 10% critique grades.

5. Final Client Project · Workforce Compensation Analytics

Students advise an executive-search steering committee using the supplied global salary archive for data, AI, and machine-learning roles. The analysis must culminate in a defensible target employer, priority roles and experience tiers, directional compensation benchmarks, and an honest statement of unsupported claims. The submission is:

Open the Gradescope assignment from Canvas and submit the single .html file by Friday, October 9 at 11:59 p.m. Eastern. Do not submit a ZIP archive, separate asset folder, or separate source file.

The companion-course presentation is not part of the 36-613 grade.

The report includes 6-10 graphics, a cohesive client story, and at least one statistical, model-based, sensitivity, or uncertainty component that complements a graph. Students must audit exact duplicates, uneven time coverage, sparse subgroups, and the absence of employer identity or row-level provenance before making market claims.

The project expectations are described in final-project.md and project-rubric.md.

Grading Policies

Final grades will be computed according to the following tentative weights:

Component Weight
Homework assignments 20%
Occasional checkpoints 20%
Graphics critique 1 10%
Graphics critique 2 10%
Final project report 40%

Final letter grades will use the following graduate plus/minus scale:

Letter grade Course average
A 93-100
A- 90 to below 93
B+ 87 to below 90
B 83 to below 87
B- 80 to below 83
C+ 77 to below 80
C 73 to below 77
C- 70 to below 73
D+ 67 to below 70
D 60 to below 67
R Below 60

MADS students must earn at least a B- for the course to count toward degree requirements. Final grades may be curved, but any curve will only improve grades relative to this scale.

Computing

All course assignments must be written in R and Quarto/R Markdown unless otherwise specified. Students should install recent versions of R and RStudio. We will primarily use tidyverse packages, especially ggplot2, dplyr, tidyr, readr, stringr, lubridate, and scales. Some topics may use additional packages for maps, text, animation, or specialized graphics.

Students are responsible for making sure their submitted files open correctly. If the submitted HTML cannot be opened or is missing embedded figures, styling, or other assets, it may lose credit.

Course Platforms and Logistics

Course Website

The public course website is the source for the syllabus, calendar, Lab 00, unannotated slides, homework instructions and files, final-project instructions and rubric, and public datasets. Assignment data are available either directly from the website or through links in the assignment.

Canvas

Canvas is the official source for announcements and summary grades. It also provides recordings, occasional checkpoints, annotated slides after the related homework is due, solutions, and the Piazza link. Solutions, annotated slides, and other Canvas-only materials are for enrolled students and may not be publicly distributed. Students are expected to check Canvas regularly.

Gradescope

Open each Gradescope assignment through its link in Canvas. Submit each homework as one PDF and the final project as one self-contained HTML file. Gradescope also contains detailed grading feedback and is the place to request regrades for homework and the final project.

Piazza

Use Piazza for class discussion, course questions, and the two graded visualization critiques: join the 36-613 Piazza.

Please allow reasonable time for a response; questions posted shortly before a deadline may not be answered in time.

Students may ask conceptual and debugging questions, but should not post complete assignment solutions, Canvas-only materials, or requests for code that solves an assignment. Submitted code and written work must be your own.

Email

Email is best for administrative and logistical questions. Questions about homework content are usually better handled in class, office hours, or the discussion board. If an email question requires a long technical explanation, I may ask that we discuss it in office hours instead.

Homework Format

Unless otherwise specified, homework should include:

Upload the PDF to Gradescope. Do not upload a ZIP archive or source file unless an assignment explicitly requests it.

You are encouraged to discuss course concepts with classmates, but the work you submit must be your own. This includes code, written explanations, and graphics.

Extensions and Late Work

In general, late assignments are not accepted unless an extension has been granted. Extensions may be granted for illness, family emergency, religious observance, interviews, or other significant circumstances. When possible, request extensions at least 48 hours before the deadline. Unexpected emergencies should be communicated as soon as reasonably possible.

Regrades

If you believe there was a grading mistake, review the assignment solutions or rubric first. Regrade requests for homework and the final project should be submitted through Gradescope, not by email, and should clearly explain the suspected grading error. Questions about checkpoint or critique grades should be submitted through Canvas. Regrade requests must be submitted within one week of grades being returned.

Academic Integrity

Cheating, plagiarism, and unauthorized assistance will be handled according to the Carnegie Mellon University academic integrity policy:

https://www.cmu.edu/academic-integrity/

Students may use course materials, documentation, and approved computing tools. Submitted work must reflect the student's own understanding and effort. If AI tools are used for brainstorming, debugging, or editing, students are responsible for verifying correctness and should follow any additional AI-use guidance provided for a specific assignment.

Disability Services

If you require accommodations, please contact the Office of Disability Resources and provide the appropriate documentation. We cannot provide formal accommodations without documentation from the university.

A Note on Diversity

My goal is for this course to support students from all backgrounds and perspectives. The diversity of the CMU community is a strength, and course materials and discussion should be respectful of differences in gender, sexuality, disability, age, race, ethnicity, nationality, culture, religion, and socioeconomic status. Suggestions for improving the course climate or accessibility are welcome.

A Note on Life-Work Balance

I care about the course material, but I care more about your well-being. Please take care of your physical and mental health, and use the resources available to you. CMU resources include:

If something serious is affecting your ability to participate in the course, please reach out sooner rather than later. We can usually do more when we know early.