Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› newms-tables-figures
newms-tables-figures
GitHub用于为《新媒体与社会》稿件设计自包含的图表、引用表和截图。确保跨学科可读性，遵循匿名化、展示不确定性及无障碍原则，并严格控制字数预算以符合期刊要求。
Trigger Scenarios
设计任何表格或图表（含定性展示和平台截图）
审稿人指出图表不清晰、冗余或缺乏说服力
在严格的字数限制下决定保留还是删减图表
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill newms-tables-figures -g -y
SKILL.md
Frontmatter
{
"name": "newms-tables-figures",
"description": "Use when building exhibits for a New Media & Society (NM&S) manuscript — quote\/excerpt tables, coding-scheme tables, network and computational figures, and screenshots of platform interfaces. Makes exhibits clear and self-contained; it does not run the analysis."
}
Tables & Figures (newms-tables-figures)
NM&S exhibits must be self-contained and legible to an interdisciplinary readership — a discourse scholar should grasp your network figure, and a computational scholar should grasp your quote table. Crucially, NM&S counts tables and figures toward the ~8,000-word target (the limit is all text including notes, references, tables, and charts), so every exhibit must earn its space.
When to trigger
- Designing any table or figure, including qualitative exhibits and platform screenshots
- A reviewer found an exhibit unclear, redundant, or unconvincing
- Deciding what to keep in-text vs. trim, given the strict word/space budget
Exhibit types and what NM&S expects
| Exhibit | Make it carry | Common failure |
|---|---|---|
| Quote / excerpt table | each quote tied to a claim + informant ID | a wall of decontextualized quotes |
| Coding-scheme table | categories, definitions, example, reliability | categories with no definitions |
| Descriptive / content-distribution table | counts + proportions + uncertainty | raw counts only |
| Network figure | what nodes/edges mean, layout logic | a hairball with no legend |
| Computational result figure | measure + validation noted, CIs | a pretty plot, no uncertainty |
| Platform screenshot / interface | annotated, anonymized, permissions noted | unredacted personal data |
Design principles (NM&S house expectations)
- Self-contained. Title, full caption, units, and source let the exhibit stand alone without the text.
- Anonymize and redact. Screenshots and quotes must not expose identifiable users, handles, faces,
or private content (see
newms-transparency-and-data). - Show uncertainty. Content distributions and computational measures carry intervals or reliability, not bare point values.
- Accessible. Colorblind-safe palettes, grayscale-legible, readable at print size; vector output.
- Earn the space. Because exhibits count toward the word target, cut any table that a sentence could replace; merge overlapping figures.
Qualitative exhibits matter at NM&S
A well-built quote-to-claim table (claim → representative excerpt → informant ID → prevalence) does real analytic work: it shows the link from evidence to argument and lets a referee audit it fast. A coding-scheme table with definitions and an example per category is often the difference between a discourse paper read as systematic vs. impressionistic.
Worked micro-example (illustrative)
Exhibit: a quote-to-claim table for "datafied control."
Columns: Claim | Representative excerpt | Informant (pseudonym, role) | # informants expressing
Rows tie each sub-claim (anticipatory compliance; score anxiety; opacity) to evidence + prevalence.
Ethics: handles removed, employer unnamed, screenshots of the app cropped to hide other users.
Space check: replaces three paragraphs of scattered quotes → saves words against the 8,000 target.
Referee pushback → NM&S-specific fix
- "Your quotes feel cherry-picked." → Convert to a quote-to-claim table with prevalence counts.
- "This figure is a hairball." → Add a legend, prune to a backbone, and state the layout algorithm.
- "You exceed the word count." → Cut redundant tables; remember exhibits count toward the target.
Calibration anchors
- Exhibits count toward the word target. Unlike journals that exclude them, NM&S includes tables and charts in the ~8,000-word total — be ruthless about which ones earn their space.
- Qualitative tables are real exhibits. A quote-to-claim or coding table is analysis made auditable.
- Anonymize every exhibit. A leaked handle or face in a screenshot is an ethics failure, not a typo.
Anti-patterns
- A wall of quotes with no claim mapping or prevalence
- Coding categories listed without definitions or examples
- Network/computational figures with no legend, units, or uncertainty
- Unredacted screenshots exposing identifiable users or private content
- Padding the paper with exhibits that push it over the word target
Output format
【Exhibit】type and the one claim it carries
【Self-contained】caption/units/source complete? [Y/N]
【Anonymized】quotes/screenshots redacted? [Y/N]
【Uncertainty】reliability/intervals shown where relevant? [Y/N]
【Space】counts toward word target — justified vs. cut?
【Next】newms-writing-style
Supplementary resources
../../resources/external_tools.md— figure, network, and CAQDAS exhibit tooling../../resources/official-source-map.md— NM&S word-target and format facts
Version History
- 1839142 Current 2026-07-05 14:07
