Exercises: Visualization Workflow
Part A: Conceptual (6 problems)
A.1 ★☆☆ | Recall
Name the 8 steps of the visualization workflow in order.
Guidance
(1) Question, (2) Data assessment, (3) Chart selection, (4) Paper sketch, (5) Prototype, (6) Critique, (7) Refine, (8) Publish.A.2 ★☆☆ | Recall
What is the chapter's threshold concept?
Guidance
"Process over product." The quality of a visualization is determined by the quality of the process that produced it. Rushing steps or skipping critique produces mediocre work regardless of technical skill.A.3 ★★☆ | Understand
Why is paper sketching before coding valuable?
Guidance
Sketching forces design decisions that code defers. On paper, you have to draw every element, which means you have to choose. Coding inherits defaults; sketching produces intended choices. A minute of sketching saves an hour of code iteration.A.4 ★★☆ | Understand
What is a critique rubric, and why should you use one instead of gut feeling?
Guidance
A rubric is a structured checklist of items to evaluate (data integrity, encoding, readability, accessibility, ethics, etc.). Using a rubric forces you to check items you might otherwise skip. Gut feeling misses systematic issues because we are blind to our own assumptions.A.5 ★★★ | Analyze
How does the workflow prevent the common failure of "rushing to code without thinking"?
Guidance
The workflow puts three steps (question, data assessment, chart selection) before code. Each takes only minutes but forces clarification of what you're doing and why. By the time you start coding, you know what you want to build. Without these steps, you discover the right answer through code iteration, which is slower and more frustrating.A.6 ★★★ | Evaluate
Can the workflow be applied at different paces for different chart stakes? Give examples.
Guidance
Yes. For a quick internal chart, the steps compress to seconds (question in one sentence, data assessment via `df.info()`, chart selection intuitive, sketch mental, prototype fast, critique brief, minimal refinement, casual publication). For a journal figure, each step gets full attention (extensive question framing, rigorous data validation, multiple sketch alternatives, careful prototype, detailed critique with external reviewers, multiple refinement rounds, polished publication). The steps are the same; only the pace varies.Part B: Applied (10 problems)
B.1 ★☆☆ | Apply
Take a dataset from a recent project and write a one-sentence question it should answer.
Guidance
Make the question specific and testable. "How has [metric] changed [time period] by [dimension]?" is a good template. Avoid vague questions like "show the data."B.2 ★☆☆ | Apply
Do a data assessment on the dataset: columns, types, row count, missing values, ranges.
Guidance
df.info()
df.describe()
df.isnull().sum()
df.head()
df.shape
B.3 ★★☆ | Apply
Pick a chart type that answers your question. Justify the choice.
Guidance
Reference Chapter 5's decision framework. Document why you picked this type over alternatives.B.4 ★★☆ | Apply
Sketch the chart on paper. Label axes, title, encoding, annotations.
Guidance
The sketch doesn't have to be artistic. It needs to show the essential structure of the intended chart. Include at least 3 alternatives if you're unsure.B.5 ★★☆ | Apply
Prototype the chart in matplotlib in 5 minutes or less.
Guidance
Focus on speed, not polish. The prototype exists to be critiqued, not to be shipped.B.6 ★★☆ | Apply
Apply the minimal critique rubric (10 items) from Section 33.7 to your prototype.
Guidance
Go through each item explicitly. Don't skip. Note which items pass and which need work.B.7 ★★★ | Apply
Refine the chart based on the critique. Document each change and why you made it.
Guidance
Each refinement should address one specific rubric failure. Don't change everything at once. Re-critique after each major change.B.8 ★★☆ | Apply
Publish the chart in an appropriate format (PNG for web, PDF for print, HTML for interactive).
Guidance
Choose based on delivery context. Save the source code alongside the output.B.9 ★★☆ | Apply
Write a one-paragraph reflection on which steps of the workflow were most valuable and which you would have skipped without discipline.
Guidance
Common answers: "I would have skipped the sketch" or "I would have skipped the critique." Noticing what you would skip helps you build the discipline to not skip.B.10 ★★★ | Create
Run a group critique session (2-6 people) on a chart using the Master Critique Rubric (Section 33.10). Document the feedback.
Guidance
Follow the rules from Section 33.15: presenter doesn't defend, critics are specific, moderator keeps things moving. Aim for 15-30 minutes per chart.Part C: Synthesis (4 problems)
C.1 ★★★ | Analyze
Reflect on a chart you've produced in the past. Which steps of the workflow did you follow implicitly? Which did you skip? What was the consequence?
Guidance
Most people skip the sketch, the critique, or both. Common consequences: chart required rework, critic caught issues the author missed, final output was inconsistent with brand. Acknowledging these helps you invest in the skipped steps in the future.C.2 ★★★ | Evaluate
Take a published chart from a news article or report. Try to infer which workflow steps the author followed.
Guidance
Clues: action title = Step 1 (clear question) was done well. Colorblind-safe palette = critique step was rigorous. Source attribution = publication checklist was followed. Missing any of these suggests the corresponding step was skipped.C.3 ★★★ | Create
Build your personal visualization toolkit document. List your go-to libraries, style sheet, helper functions, critique rubric, and references.
Guidance
Follow Section 33.11 as a template. About 1-2 pages. Keep it updated as your practice evolves.C.4 ★★★ | Evaluate
The chapter's threshold concept is "process over product." Do you agree? Could someone produce great charts without following a process?
Guidance
Experienced practitioners appear to produce great charts quickly, but they are actually following the process rapidly, not skipping it. A true beginner with no process produces inconsistent results — sometimes good, often bad. The process is what makes quality reliable rather than accidental.Chapter 34 is the capstone project: a complete end-to-end data story applying all the workflow steps and techniques from the book.