Key Takeaways: Visualization Workflow

  1. Process over product. The chapter's threshold concept: the quality of a visualization is determined by the quality of the process that produced it. Technical skill is necessary but not sufficient; a disciplined process is what turns skill into consistent output.

  2. The 8-step workflow. Question → Data → Chart type → Sketch → Prototype → Critique → Refine → Publish. Each step has specific goals, deliverables, and common failure modes. Apply them in order, iterating when critique reveals issues.

  3. Question before code. Every visualization starts with a specific, testable question. Not a dataset, not a tool — a question. Write it in one sentence before doing anything else. Multiple questions in one chart is the most common failure at Step 1.

  4. Data assessment catches problems early. Before designing a chart, understand the data: columns, types, ranges, missing values, provenance. Five minutes of df.info() and df.describe() prevents hours of debugging later.

  5. Paper sketching beats code iteration. Sketching on paper forces design decisions that coding inherits from defaults. A minute of sketching saves an hour of coding. Consider 2-3 alternatives before picking one.

  6. Prototype fast, polish later. The prototype exists to be critiqued, not shipped. Over-polishing before critique wastes time if the critique reveals a wrong approach. Speed over quality at this stage.

  7. Critique with a rubric, not gut feeling. A structured rubric forces you to check specific items (data integrity, encoding, readability, accessibility, ethics, narrative). Self-critique is valuable; group critique is better. Skipping critique is the most common failure in the workflow.

  8. Refine iteratively; know when to stop. Each refinement addresses one rubric failure. Re-critique after major changes. Stop when the chart answers the question clearly and passes the rubric. Over-refining is a waste of time.

  9. Publishing is a complete handoff. Format selection, DPI, branding, caption, distribution, and archiving. Every chart, every time. The publish step is not just "save the file."

  10. Build a personal toolkit. Go-to libraries, style sheets, helper modules, critique rubrics, example galleries. The toolkit is the physical embodiment of the workflow. Grow it over time; keep it lean enough to use daily.

Chapter 34 is the book's capstone project: a complete end-to-end data story using the full 8-step workflow and every technique from the book. Chapter 35 is the reference gallery of example charts and patterns that serves as a permanent lookup resource.