Design reading notes

Stephen Few / Perceptual Edge — full legal white papers

Stephen Few / Perceptual Edge — full legal white papers

Legal source: https://www.perceptualedge.com/library.php (author-hosted white papers / PDFs). Read in full via official PDF URLs. These sit in the same design lineage as Tufte (clarity, remove chartjunk, accurate encoding).

---

1. Common Pitfalls in Dashboard Design

1. Exceeding a single screen (scrolling/fragmentation destroys simultaneity)

2. Inadequate context for measures (compared to what? good/bad?)

3. Excessive detail or precision

4. Expressing measures indirectly (show variance/percentages when that’s the need)

5. Inappropriate display media (wrong chart type; pie when bars work)

6. Meaningless variety

7. Poorly designed display media

8. Encoding quantitative data inaccurately

9. Poor arrangement of data

10. Ineffective highlighting of what’s important

11. Cluttering with useless decoration

12. Misusing or overusing color

13. Unappealing visual display

2. Effectively Communicating Numbers: Selecting the Best Means and Manner of Display

- Graphicacy is learned, not intuitive; most business graphs miscommunicate unintentionally.

- Always ask “Compared to what?”—numbers gain meaning in relation.

- Tables vs graphs: tables for lookup/precise values; graphs for shape (patterns, trends, exceptions).

- Quantitative vs categorical data; categorical scales: nominal, ordinal, interval.

- Seven common relationships: time-series, ranking, part-to-whole, deviation, distribution, correlation, nominal comparison.

- Best quantitative encodings: length and 2-D position (points, lines, bars, boxes); avoid 2-D area (pies/bubbles) when precision matters.

- Lines only along interval-like scales; bars require a zero baseline (else use points).

- Remove anything that doesn’t contribute meaning (distraction).

- Six-stage design process: message/data → table vs graph → encoding → variable placement → remaining object design → featuring important data.

3. Visual Communication: Core Design Principles for Displaying Quantitative Information

1. Display neither more nor less than what is relevant to the message (no chartjunk / no missing essentials).

2. Do not include visual differences that do not correspond to data differences.

3. Prefer length or 2-D location to encode quantities (preattentive attributes; size/intensity only for rough approximations).

4–7. (Continuing perceptual rules in the paper:) design for how people see and think; match message first; use perception-based rules so important patterns stand out—otherwise displays mislead or become incomprehensible.

---

Also located (not required beyond 2–3; not fully logged here)

Library lists additional legal white papers (e.g., Dashboard Design for Real-Time Situation Awareness). Three full papers above meet the 2–3 white-paper target.

---

Not obtained