assess-graphical-excellence

Name: assess-graphical-excellence
Author: gnurio

6installs18forks

How to install assess-graphical-excellence?

gnurio/tufte-vdqi-plugin/assess-graphical-excellence

$npx -y skills add gnurio/tufte-vdqi-plugin --skill assess-graphical-excellence

Installs into the current project.

Summary

Evaluate a data graphic against Edward Tufte's nine criteria, name the chartjunk species present, compute lie factor, compare against VDQI's named-failure catalogue, and return prioritised fixes tagged with the specific Tufte remedy (B1–B7), genre to switch to (C1–C10), and exemp

Files · 1

View on GitHub

SKILL.md

1	# Assess Graphical Excellence
2
3	This is the assessment hub of the Tufte toolkit. Its job is to score an existing
4	graphic, name what's wrong using Tufte's vocabulary (the duck, the dreaded
5	grid, moiré vibration, dimensionality violation, etc.), and hand back the exact
6	remedy plus the genre to switch to and an exemplar to emulate.
7
8	The output is concrete and source-grounded because the model reasons over a
9	single principles file (`references/tufte-principles.md`) that quotes VDQI by
10	page. Generic "improve data-ink ratio" advice is the failure mode this skill is
11	designed to avoid.
12
13	## What you need from the user
14
15	A description (or image, or file) of the graphic, and ideally its purpose and
16	audience. If something essential is missing, infer reasonably and state the
17	assumption rather than stalling.
18
19	## How to assess (six-step workflow)
20
21	Read `references/tufte-principles.md` first. The workflow uses every part of it.
22
23	1. Score the nine criteria (Part A). 0–10 each, with a chart-specific
24	observation. Use VDQI's numeric anchors: e.g. data-ink ratio "0.1–0.2 is
25	typical, edit toward 1.0" (p.136); data density 0.15 numbers/in² is
26	"overwrought" (p.162). Unsupported scores are the main failure mode.
27
28	2. Compute the lie factor when proportionality looks suspicious (Part B,
29	B1). Formula: `(visual change %) / (data change %)`. Acceptable 0.95–1.05.
30	Report the number and compare to VDQI's catalogue (Part E) — anchor the
31	verdict in a named case: "this is essentially the 1979 TIME barrel
32	(lie factor 59.4)" or "in the league of the LA Times shrinking-doctor
33	(lie factor 2.8)" rather than a free-floating number.
34
35	3. Identify chartjunk species present (Part D). Walk the four named
36	offenses and name each one the chart exhibits, citing Tufte's named offender
37	when the resemblance is close:
38	- Moiré vibration (cross-hatching, dense stippling, gradients)
39	- Dreaded grid (grid darker than the data marks)
40	- Duck (decoration drives the chart; visual style > data; dimensionality
41	exceeds data dimensionality)
42	- Decoration (ornament that carries no information)
43
44	4. Rank candidate genres and challenge the default (Part C, Part G). List
45	at least three Tufte genres that could fit the data, ranked by fit. Then
46	apply the default-challenge rule:
47
48	> If your top-ranked genre is also what an unprompted Claude would pick
49	> (line, bar, dot plot, scatter, pie), you MUST do one of:
50	> (a) explicitly justify it by citing what the alternatives lose, OR
51	> (b) reach for a second-line VDQI move (supertable, table-graphic,
52	> sparkline, dot-dash plot, quartile plot, small multiples) and explain
53	> why it is the stronger fit here.
54	>
55	> Quiet defaulting to the obvious chart is the failure mode this rule
56	> exists to catch.
57
58	Common multi-answer data shapes (flag these in the output so render knows
59	to emit alternatives):
60	- 1 number / single ratio → prose statement + tiny inline visual
61	- ≤20 numbers → supertable + Tufte chart (VDQI p.56)
62	- Many series of one x → small multiples + overplotted comparison
63	- Distributions across groups → quartile plot + strip plot or histogram
64	- Bivariate scatter → range frame + dot-dash marginal variant
65
66	5. Compute the weighted overall score. Weights: integrity 3×,
67	proportionality 2×, data-ink 2×, typography 0.5×, the rest 1×.
68
69	6. Translate scores into ranked fixes. Each fix gets up to four tags:
70	- Remedy: B1–B7 (the technique).
71	- Genre: C1–C10 (the form to switch to). Optional.
72	- Anti-pattern resemblance: name a Part-E case the graphic looks like.
73	Optional.
74	- Exemplar to emulate: name a Part-F graphic the redesign should
75	resemble. Optional.
76
77	If the graphic plots a multi-year currency series, check B7. Use
78	`scripts/deflate.py` (requires real CPI values; refuses to guess).
79
80	## Output format
81
82	```
83	## Assessment: <graphic>
84	Context: <purpose / audience, or stated assumption>
85
86	### Scores
87	<one line per criterion: name — score/10 — chart-specific observation>
88
89	### Chartjunk species present
90	<list any of: moiré, dreaded grid, duck, decoration — each with the detected signature, and Tufte's named offender it most resembles>
91
92	### Distortion check
93	Lie factor: <value or "n/a"> — <interpretation>
94	Resembles: <named VDQI case from Part E, or "no close analogue">
95
96	### Genres considered (ra

People who liked this also viewed

6 picks

Type

assess-graphical-excellence

6installs18forks

View on GitHub

How to install assess-graphical-excellence?

gnurio/tufte-vdqi-plugin/assess-graphical-excellence

$npx -y skills add gnurio/tufte-vdqi-plugin --skill assess-graphical-excellence

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Assess Graphical Excellence
2
3	This is the assessment hub of the Tufte toolkit. Its job is to score an existing
4	graphic, name what's wrong using Tufte's vocabulary (the duck, the dreaded
5	grid, moiré vibration, dimensionality violation, etc.), and hand back the exact
6	remedy plus the genre to switch to and an exemplar to emulate.
7
8	The output is concrete and source-grounded because the model reasons over a
9	single principles file (`references/tufte-principles.md`) that quotes VDQI by
10	page. Generic "improve data-ink ratio" advice is the failure mode this skill is
11	designed to avoid.
12
13	## What you need from the user
14
15	A description (or image, or file) of the graphic, and ideally its purpose and
16	audience. If something essential is missing, infer reasonably and state the
17	assumption rather than stalling.
18
19	## How to assess (six-step workflow)
20
21	Read `references/tufte-principles.md` first. The workflow uses every part of it.
22
23	1. Score the nine criteria (Part A). 0–10 each, with a chart-specific
24	observation. Use VDQI's numeric anchors: e.g. data-ink ratio "0.1–0.2 is
25	typical, edit toward 1.0" (p.136); data density 0.15 numbers/in² is
26	"overwrought" (p.162). Unsupported scores are the main failure mode.
27
28	2. Compute the lie factor when proportionality looks suspicious (Part B,
29	B1). Formula: `(visual change %) / (data change %)`. Acceptable 0.95–1.05.
30	Report the number and compare to VDQI's catalogue (Part E) — anchor the
31	verdict in a named case: "this is essentially the 1979 TIME barrel
32	(lie factor 59.4)" or "in the league of the LA Times shrinking-doctor
33	(lie factor 2.8)" rather than a free-floating number.
34
35	3. Identify chartjunk species present (Part D). Walk the four named
36	offenses and name each one the chart exhibits, citing Tufte's named offender
37	when the resemblance is close:
38	- Moiré vibration (cross-hatching, dense stippling, gradients)
39	- Dreaded grid (grid darker than the data marks)
40	- Duck (decoration drives the chart; visual style > data; dimensionality
41	exceeds data dimensionality)
42	- Decoration (ornament that carries no information)
43
44	4. Rank candidate genres and challenge the default (Part C, Part G). List
45	at least three Tufte genres that could fit the data, ranked by fit. Then
46	apply the default-challenge rule:
47
48	> If your top-ranked genre is also what an unprompted Claude would pick
49	> (line, bar, dot plot, scatter, pie), you MUST do one of:
50	> (a) explicitly justify it by citing what the alternatives lose, OR
51	> (b) reach for a second-line VDQI move (supertable, table-graphic,
52	> sparkline, dot-dash plot, quartile plot, small multiples) and explain
53	> why it is the stronger fit here.
54	>
55	> Quiet defaulting to the obvious chart is the failure mode this rule
56	> exists to catch.
57
58	Common multi-answer data shapes (flag these in the output so render knows
59	to emit alternatives):
60	- 1 number / single ratio → prose statement + tiny inline visual
61	- ≤20 numbers → supertable + Tufte chart (VDQI p.56)
62	- Many series of one x → small multiples + overplotted comparison
63	- Distributions across groups → quartile plot + strip plot or histogram
64	- Bivariate scatter → range frame + dot-dash marginal variant
65
66	5. Compute the weighted overall score. Weights: integrity 3×,
67	proportionality 2×, data-ink 2×, typography 0.5×, the rest 1×.
68
69	6. Translate scores into ranked fixes. Each fix gets up to four tags:
70	- Remedy: B1–B7 (the technique).
71	- Genre: C1–C10 (the form to switch to). Optional.
72	- Anti-pattern resemblance: name a Part-E case the graphic looks like.
73	Optional.
74	- Exemplar to emulate: name a Part-F graphic the redesign should
75	resemble. Optional.
76
77	If the graphic plots a multi-year currency series, check B7. Use
78	`scripts/deflate.py` (requires real CPI values; refuses to guess).
79
80	## Output format
81
82	```
83	## Assessment: <graphic>
84	Context: <purpose / audience, or stated assumption>
85
86	### Scores
87	<one line per criterion: name — score/10 — chart-specific observation>
88
89	### Chartjunk species present
90	<list any of: moiré, dreaded grid, duck, decoration — each with the detected signature, and Tufte's named offender it most resembles>
91
92	### Distortion check
93	Lie factor: <value or "n/a"> — <interpretation>
94	Resembles: <named VDQI case from Part E, or "no close analogue">
95
96	### Genres considered (ra

People who liked this also viewed

6 picks

Type