assets/templates/editorial-evaluation-rubric.md

# Editorial Visualization Evaluation Rubric

Score each category from 1 to 5. A production-ready editorial figure should usually score 28 or higher out of 35, with no category below 4 unless the tradeoff is deliberate and documented.

| Category | 1 | 3 | 5 | Score |
|---|---|---|---|---|
| Clarity | Main point is hard to find. | Main point is visible after some decoding. | Main comparison is immediately readable. |  |
| Storytelling strength | Chart names data but makes no claim. | Figure suggests a claim but reading path is loose. | Title, evidence, and annotations form a clear argument. |  |
| Hierarchy | Every element has similar weight. | Some priority is visible but controls or legends compete. | Eye lands on the key claim, then evidence, then context. |  |
| Restraint | Color, grid, decoration, or controls distract. | Some nonessential ink remains. | Every visible element has a job. |  |
| Annotation quality | Annotations restate values or clutter the chart. | Some notes explain significance. | Notes explain turning points, mechanisms, caveats, or consequences. |  |
| Accessibility | Depends on color, hover, tiny text, permissions, or pointer precision. | Has partial labels or alt text but gaps remain. | Works in grayscale, narrow width, touch/focus paths, export, and accessible summary. |  |
| Originality | Copies a known publication style or generic template. | Adapted template with some story-specific choices. | Composition is specific to the data and audience without copying a brand. |  |

## Editorial Ambition Checks

These do not change the 35-point score, but they determine whether the work is merely polished or genuinely editorial.

| Check | Pass Criteria | Notes |
| --- | --- | --- |
| Artifact choice | The team chose data-first chart, generated object marks, illustrated substrate, flow map, 3D surface, or scrollytelling before choosing the renderer. |  |
| Image usefulness | Any generated imagery or illustration carries scale, place, mechanism, motion, or stakes. |  |
| Motion usefulness | Every animation has a named explanatory verb and a reduced-motion fallback. |  |
| Human edit | Labels, omissions, hierarchy, and pacing feel deliberately edited rather than template-produced. |  |
| Still-frame strength | The first frame, final frame, and a screenshot each preserve the main claim. |  |
| Mobile concept set | Large-screen and mobile portrait concepts were generated and approved; mobile landscape exists when justified. |  |
| Mobile interaction resilience | Mobile keeps the main visualization visible, handles touch/keyboard/spotty connection, and uses device capabilities only with fallbacks. |  |
| Fictional data richness | Fictional or synthetic stories include entity, temporal, spatial/physical, event, outcome, and derived comparison layers with a seed or regeneration path. |  |
| Embedded visualization self-use | Each embedded chart, map, table-graphic, swarm, distribution, flow layer, particle layer, media overlay, key, or fallback has a specialist owner, mini-brief, QA check, and delegated or local fresh specialist pass before integration. |  |

## Sensitive Subject Checks

Use these for conflict, disaster, displacement, civilian harm, migration, political violence, or humanitarian work. A failure here should block publication even if the 35-point score is high.

| Check | Pass Criteria | Notes |
| --- | --- | --- |
| Source hierarchy | Map, human-impact, contextual, and research layers have visible sources and method notes. |  |
| Dated evidence | Temporal map states, event values, and estimates show the date or update cadence readers need. |  |
| Evidence status | Measured, estimated, schematic, and disputed layers are distinguishable in labels and styling. |  |
| Humane framing | Language and visual effects preserve dignity and avoid spectacle, decorative violence, or team-color framing. |  |
| Static accountability | A screenshot or export preserves claim, caveat, source context, and human-impact framing without hover or autoplay. |  |

## Required Review Notes

- Takeaway:
- Evidence that supports it:
- Artifact mode:
- Image or illustration role:
- Motion role:
- Fictional data richness:
- Embedded visualization self-use notes:
- Large-screen concept:
- Mobile portrait concept:
- Mobile landscape concept, if needed:
- Strongest annotation:
- Weakest hierarchy issue:
- Mobile/narrow risk:
- Touch, keyboard, spotty-connection, or permission risk:
- Accessibility risk:
- Originality check:
- Sensitive-subject risk:
- Decision: ship, revise, or reframe.