Data Visualization Storytelling — Visual Narrative Design

Designs visual narratives that transform data into comprehension through chart selection, Mermaid diagram metodologia-storytelling, annotation strategy, and dashboard sequencing. Owns the visual layer of data communication across all discovery deliverables.

Guiding Principle

A visualization that requires explanation has failed. The right diagram, with the right annotations, tells the story on its own. Text complements — it does not replace. Every visualization has ONE main message, and all visual design directs attention to that message.

Visual Philosophy

1. One message per visualization. If you need to communicate two things, use two visualizations.
2. Form follows data. Comparison → bars. Trend → line. Composition → stacked. Relationship → scatter. Flow → Mermaid.
3. Selective annotation. Annotate what matters, not everything. Excessive annotations equal zero annotations.
4. Narrative sequence. Visualizations in a document/presentation build an argument, they are not standalone decorations.

Inputs

• $1 — Visualization context: analysis, presentation, dashboard, comparison, flow (default: analysis)
• $2 — Format target: markdown, html, pptx (default: markdown)

Parse from $ARGUMENTS.

Chart Selection Matrix

Mermaid Narrative Design

Every Diagram Tells a Story

BAD: Diagram shows components and connections
  → Reader: "OK, there are boxes and arrows. So what?"

GOOD: Diagram shows WHY this architecture matters
  → Title: "Pain point: 3 services share 1 database"
  → Highlighted node: the shared database (red classDef)
  → Annotation: "Single point of failure — all 3 services go down together"
  → Narrative text before: "El acoplamiento en la capa de datos..."

Mermaid Standards

Color Strategy (MetodologIA Brand)

classDef primary fill:#6366F1,stroke:#1A1A2E,color:#fff    %% MetodologIA orange — key components
classDef risk fill:#DC3545,stroke:#1A1A2E,color:#fff       %% Red — risk/problem areas
classDef success fill:#22D3EE,stroke:#1A1A2E,color:#1A1A2E %% Gold — success (NEVER green)
classDef neutral fill:#F8F9FA,stroke:#1A1A2E,color:#1A1A2E %% Light — supporting components

Annotation Strategy

What to Annotate

What NOT to Annotate

• Every data point (clutters)
• Obvious patterns (reader can see)
• Units/labels already in axes (redundant)
• Decorative elements (distracting)

Visual Hierarchy in Documents

Per-Deliverable Diagram Budget

Visual Sequence in Presentations (PPTX)

Slide 1: The headline visual (single powerful chart/diagram)
Slide 2: The context visual (trend, timeline, or process)
Slide 3: The evidence visual (detailed comparison or matrix)
Slide 4: The action visual (roadmap, decision tree, or next steps)

Each visual builds the argument. No decorative slides.

Format-Specific Guidelines

Markdown

• Fenced \``mermaid` blocks
• Text summary BEFORE diagram (accessibility)
• Source tag AFTER diagram

HTML

• <pre class="mermaid"> with CDN v10
• Responsive sizing
• Print-ready @media print fallback

PPTX

• Pre-rendered Mermaid as images
• One visual per slide maximum
• Speaker notes reference the data source

Output Configuration

• Language: Spanish (Latin American, business register — simple, clear, concise, direct)
• Attribution: Expert committee of the MetodologIA Discovery Framework
• Tagline: "Construido por profesionales, potenciado por la red agéntica de MetodologIA."

Validation Gate

Edge Cases

• No quantitative data for charts: Use structural Mermaid diagrams (C4, flowcharts) to tell the architecture story.
• Too many data points: Aggregate or filter. Show the top 5 + "others". Never plot 50 bars.
• Mermaid rendering limitations: Fall back to structured tables when diagrams would be illegible.

Limits

• This skill owns visual design and diagram narratives. It does NOT own metric interpretation (that's metodologia-data-storytelling) or format production (that's metodologia-output-engineering).
• Mermaid is the primary diagramming tool. No external tools or image generation.
• Follow Mermaid Diagramming Standard in CLAUDE.md as baseline.

Casos Borde

Decisiones y Trade-offs

Knowledge Graph

graph TD
    subgraph Core["Data Viz Storytelling"]
        A["Chart Selection Matrix"] --> B["Mermaid Narrative Design"]
        A --> C["Annotation Strategy"]
        B --> D["Visual Sequence"]
        C --> D
    end
    subgraph Inputs["Inputs"]
        E["Data Pattern"] --> A
        F["Visualization Context"] --> A
        G["Format Target"] --> B
    end
    subgraph Outputs["Outputs"]
        D --> H["Annotated Diagrams"]
        D --> I["Dashboard Narratives"]
    end
    subgraph Related["Related Skills"]
        J["mermaid-diagramming"] -.-> B
        K["output-engineering"] -.-> D
    end

Output Templates

Markdown (default)

• Filename: {fase}_DataViz_{cliente}_{WIP}.md
• Structure: TL;DR > Chart Selection Rationale > Mermaid diagrams with accessibility text > Annotation notes > Visual sequence narrative > ghost menu

PPTX

• Filename: {fase}_DataViz_{cliente}_{WIP}.pptx
• Structure: 1 visual per slide; speaker notes with data source and annotation rationale; narrative arc (headline visual > context > evidence > action)

HTML (bajo demanda)

• Filename: {fase}_DataViz_{cliente}_{WIP}.html
• Estructura: HTML self-contained branded (Design System MetodologIA v5). Light-First Technical. Incluye diagramas Mermaid renderizados vía CDN, anotaciones interactivas por chart y texto de accesibilidad. WCAG AA, responsive, print-ready.

DOCX (circulación formal)

• Filename: {fase}_{entregable}_{cliente}_{WIP}.docx
• Generado via python-docx con Metodología Design System v5. Portada con metadata del engagement, TOC automático, encabezados/pies de página con marca. Tablas con zebra striping, tipografía Poppins en headings (navy), Montserrat en cuerpo, acentos dorados. Para circulación formal y auditoría.

XLSX (bajo demanda)

• Filename: {fase}_{entregable}_{cliente}_{WIP}.xlsx
• Via openpyxl con MetodologIA Design System v5. Headers con fondo navy y tipografía Poppins en blanco, conditional formatting por tipo de chart y prioridad visual, auto-filters en todas las columnas, valores directos sin fórmulas.

Evaluacion

Umbral minimo: 7/10 en cada dimension para considerar el skill production-ready.