data-dashboard-intelligence_skill

Overview

This skill converts normalized event data into premium, story-driven dashboard specifications and edits. It focuses on field semantics and shape to pick components, aggregations, headlines, and layout so dashboards communicate a clear conclusion. Use it whenever you generate or modify dashboards to ensure readability, correct aggregations, and graceful handling of sparse or missing data.

How this skill works

The skill inspects each field's shape (numeric, categorical, temporal, text) and semantic hints (duration, cost, status, id) to select components and aggregations. It applies override and fallback rules to avoid broken charts, chooses a hero stat based on domain signals, and organizes components into a progressive-reveal story layout. It also enforces data-quality thresholds and graceful degradation chains so visualizations remain meaningful at any sample size.

When to use it

• Generating a new dashboard spec from normalized event fields
• Editing component types, titles, or aggregations on an existing dashboard
• Choosing chart types and time intervals based on field shape and timespan
• Selecting the hero metric and arranging KPI/supporting cards
• Handling sparse, empty, or low-quality data with fallbacks

Best practices

• Map by field semantics, not raw names: let 'duration_ms' be duration, not just 'ms'
• Pick aggregations that make sense: durations → average, money → sum, ids → count
• Place the hero stat top-left, timeseries full-width, breakdowns in the middle, table last
• Aim for 7–30 x-axis points by selecting hourly/daily/weekly/monthly appropriately
• Use clear, non-technical headlines: 'Total Executions', 'Avg Call Duration', 'Calls by Status'
• Apply graceful degradation: timeseries → bar → metric; pie → bar → metric

Example use cases

• Create a one-page dashboard for a workflow system highlighting success rate as the hero
• Convert sparse event logs into a compact KPI row + explanatory message instead of a broken chart
• Swap a pie chart for a horizontal bar when a category exceeds six values
• Detect a duration field and show Avg Call Duration rather than a misleading total
• Auto-generate component titles that replace platform terms with universal vocabulary