/chart-pack — Chart Pack

Category: Documentation · Primary agent: visualization-dashboard-specialist

Produces a chart pack from code: fan charts, shock decompositions, scenario comparisons, with consistent style and labelling.

Arguments: $ARGUMENTS

Purpose

Produce a chart pack for a forecast round, brief or presentation, generated entirely from code: fan charts, shock decompositions, scenario comparisons and supporting charts, each labeled with units, sources and vintage, accessible and consistent in style. The output regenerates from one command when data change.

Use it when model results must be communicated visually to decision-makers or in publications.

Do not use it when an interactive monitoring tool is needed rather than static charts; design a dashboard with the visualization specialist.

Before you start

1. Check required inputs. For each input below record FOUND or ABSENT (never 'assumed'):
   • Results files If a required input is ABSENT, that part is NOT ASSESSED — NO DATA: do not estimate or invent it; name what is missing and which skill produces it. If all required inputs are absent, stop and say so.
2. Check confidentiality. Identify the data classification level of the inputs (.claude/docs/confidentiality-and-ai-use.md). Do not proceed with restricted or confidential data in a way the policy does not allow.
3. Collaboration. Follow .claude/docs/collaboration-protocol.md: ask questions first, present options, get approval before writing files, and update production/session-state/active.md after each approved section.

Agents

• visualization-dashboard-specialist — primary agent; coordinates the work and integrates results
• forecast-evaluation-specialist — Evaluates forecasts

Run these agents as subagents (the Task tool), in parallel where the steps are independent. Subagents only write to the paths named in your instructions.

Key design decisions (present options, let the user choose)

Decision	Typical options	What it changes
Chart list	one chart per message	Focus
Uncertainty	fan charts, bands, scenario lines	Honesty
Format	publication, slides, web	Sizes and fonts
Style	studio template, institutional template	Consistency

Workflow

1. Choose a chart list matching the message
2. Generate charts from scripts (never by hand), with units, sources and vintage in the footer
3. Check accessibility (colors, fonts) and consistency
   • Color-blind-safe palettes; avoid dual axes and truncated axes that mislead
4. Archive scripts and outputs together

Checkpoints with the user

• After the chart list: confirm messages
• After drafts: review readability

Common pitfalls

• Too many series per chart
• Forecast charts without uncertainty
• Hand-edited charts that cannot be updated

Output

docs/briefs/<name>-charts/; a Word file with the charts and their captions on request (python tools/report.py docx <index.md> --out <file.docx>)

The folder holds the scripts, the data behind each chart and the rendered charts.

Gate / verdict

• Charts regenerate from one command; uncertainty is visible
• Every chart shows units, sources and vintage

Suggested next steps

• /policy-brief