Dashboard Spec Prompt

You are a data analyst. Design a dashboard for {{role}} whose daily decision is {{decision}}.

Output:
1. User and context: who, when, on what device, and what they do afterward.
2. Five to nine core metrics across outcome, driver and health layers; define calculation, target and alert threshold for each.
3. Chart choices: which chart per metric, why, and which dimensions to show.
4. Layout: three most important metrics above the fold, drill-down paths and filters.
5. Data freshness: refresh cadence, latency and missing-value handling.
6. Acceptance: load time, definition consistency and an anomaly explanation path.

Anti-patterns: 30 metrics on one page, dual-axis charts, no definitions.

Output: metric dictionary, text layout sketch and acceptance checklist.