菜单
Guided requirements template for data storytelling projects -- walks users through audience, Big Idea, questions, data sources, and delivery target before any chart is created
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/fabioc-aloha/Alex_ACT_Visual_Storytelling --skill storytelling-requirements命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Clear documentation through visual excellence
Data cleaning, profiling, transformation, and quality gates -- prepares raw data for visualization and analysis
Ingestion patterns for CSV, JSON, REST API, SQL, Excel, and Parquet -- guides an LLM through loading data from any common source
正在显示 SKILL.md
| type | skill |
| lifecycle | stable |
| inheritance | inheritable |
| name | storytelling-requirements |
| description | Guided requirements template for data storytelling projects -- walks users through audience, Big Idea, questions, data sources, and delivery target before any chart is created |
| tier | standard |
| applyTo | **/*story*,**/*requirements*,**/*brief*,**/*dashboard*,**/*report* |
| currency | 2026-08-15T00:00:00.000Z |
| lastReviewed | 2026-08-15T00:00:00.000Z |
Every data storytelling project starts with a brief. This plugin provides the template and guides the user through it. No chart should be created before this document is filled out.
The brief is the contract between the analyst and the audience. It answers: who is reading this, what decision does it support, and what does the data need to show?
chart-vocabulary or any delivery workflowCopy this template to your project and fill it out with the user. Each section maps to a step in the storytelling pipeline.
| Field | Answer |
|---|---|
| Primary audience | (e.g., "CFO and finance leadership team") |
| Audience expertise | Executive / Manager / Analyst / General |
| Time budget | 30 seconds / 2 minutes / Unlimited |
| Decision this supports | (e.g., "Whether to increase Q4 marketing spend") |
| How they will consume it | Screen / Print / Presentation / Email / Mobile |
Write one sentence that captures the argument:
[Audience] should [action/decision] because [evidence from data].
Example: "The exec team should double Q4 email marketing spend because every $1 generates $3.20 in pipeline."
If you cannot write this sentence, the analysis is not done yet. Go back to exploration.
Before accepting a Big Idea or decision sentence, make its evidence executable. Record one row for every decision-bearing claim:
| Claim | Required fields | Formula | Grain | Status |
|---|---|---|---|---|
| (e.g., "Email returns $3.20 per $1") | pipeline, campaign spend | pipeline / spend | campaign-quarter | Computable / Missing input |
Rules:
List 3-5 questions, ranked by priority. Each question becomes one visual.
| # | Question | Communication Goal | Priority |
|---|---|---|---|
| 1 | (e.g., "How has revenue trended quarterly?") | Change Over Time | Must-have |
| 2 | (e.g., "Which regions contribute most?") | Comparison | Must-have |
| 3 | (e.g., "What share does each product line hold?") | Proportion | Should-have |
| 4 | |||
| 5 |
Communication goals map to chart types through Illustrator's chart-vocabulary.
| Source | Format | Location | Refresh Cadence | Notes |
|---|---|---|---|---|
| (e.g., "Sales CRM export") | CSV | SharePoint folder | Monthly | Filter: last 8 quarters |
| Concern | Impact | Mitigation |
|---|---|---|
| (e.g., "Missing region codes for 12% of records") | Proportions will undercount | Map nulls to "Unknown" region |
| Field | Answer |
|---|---|
| Output format | SVG in Markdown / HTML dashboard / Power BI report / Slides / Email |
| Branding | Dark slate / Light / Custom palette |
| Interactivity | Static / Filters / Drill-through / Cross-filter |
| Hosting | GitHub README / SharePoint / Power BI Service / Local file |
| Update frequency | One-time / Weekly refresh / Real-time |
| Constraint | Detail |
|---|---|
| 5-visual rule | Executive audience: max 5 visuals per page (3 KPIs + hero + supporting) |
| Accessibility | WCAG 2.1 AA: 4.5:1 contrast, no color-only encoding |
| Token budget | (if rendering in AI context: max tokens for the output artifact) |
| Deadline | (when is this needed?) |
When a user says "build me a dashboard" or "make a report from this data", walk them through the template:
Start with Section 2 (Big Idea). If they can write the sentence, the rest flows. If they cannot, they need exploration first -- invoke data-analysis or data-storytelling Module 1 (Discover).
Fill Section 1 (Audience). This determines the 5-visual rule budget, chart complexity, and delivery format.
Fill Section 3 (Questions). Map each question to a communication goal using the heuristic table below. This is where Illustrator's chart-vocabulary fires.
Fill Sections 4-5 (Data). This determines which datasource-connectors and data-preparation steps are needed.
Fill Sections 6-7 (Delivery). Use delivery-ascii-dashboard for ASCII. For graphical output, require alex-act-illustrator-plugin.
When users struggle to name the communication goal for a question, use these trigger words:
| Trigger Words in the Question | Communication Goal |
|---|---|
| "rank", "compare", "versus", "top", "best/worst" | Comparison |
| "over time", "trend", "growth", "decline", "monthly" | Change Over Time |
| "share", "proportion", "breakdown", "percent of" | Proportion |
| "spread", "range", "outlier", "distribution" | Distribution |
| "correlation", "relationship", "predict", "affects" | Relationship |
| "flow", "path", "conversion", "funnel", "from X to Y" | Flow |
| "deviation", "variance", "above/below target" | Deviation |
After filling the brief and before building, run one CSAR pass:
In this collection, CSAR always means Clarify, Summarize, Act, Reflect. It is a review workflow, not a substitute for the Claim Computability Gate above.
storytelling-requirements (this plugin -- the brief)
|
v
datasource-connectors (ingest data per Section 4)
|
v
data-preparation (clean per Section 5 concerns)
|
v
Illustrator chart-vocabulary (select charts per Section 3 goals)
|
v
ASCII delivery or Illustrator (render per Section 6 target)
| Anti-pattern | Fix |
|---|---|
| Building charts before writing the Big Idea | The Big Idea is the filter. Without it, you build 20 charts and keep 5. |
| Skipping the audience section | Executive dashboards and analyst reports are different artifacts. The audience determines everything. |
| Listing 10+ questions for an executive dashboard | 5-visual rule. Cut to 3-5 questions or split into multiple pages. |
| No data quality section | Surprises in the data surface during the demo, not the build. Document concerns early. |
| Filling the template alone (without the stakeholder) | The brief is a conversation tool. Fill it together. |
chart-vocabulary from alex-act-illustrator-plugin -- chart selection by communication goal (fires after Section 3)data-storytelling -- narrative arc and orchestration (fires after the brief is complete)data-visualization -- chart design: color, annotation, decluttering (fires after chart selection)data-preparation -- cleaning and profiling (fires after Section 4-5)datasource-connectors -- data ingestion (fires after Section 4)delivery-ascii-dashboard or alex-act-illustrator-plugin -- rendering after Section 6