All skills
nimrodfisher avatar

/dashboard-specification

@fcb0454

Design specifications for effective dashboards. Use when planning new dashboards, improving existing ones, or documenting dashboard requirements before development starts.

Use this Skill: https://skilld.dev/gh/nimrodfisher/data-analytics-skills/dashboard-specification

This session only. Nothing lands on disk.

referencesdashboard_design_principles.md

≈744 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Dashboard Design Principles

The single question test

Every dashboard should answer one primary question. If you cannot complete the sentence "This dashboard tells me ___," the scope is too broad. A dashboard that tries to answer five questions answers none of them well.


Information hierarchy

Level 1 — KPIs: The 2–4 metrics that matter most. Always visible above the fold. No interaction required to see them.

Level 2 — Context: Trend lines, period comparisons, variance from target. Explains whether KPIs are good or bad.

Level 3 — Diagnostic: Breakdowns by segment, time, or dimension. Used to investigate anomalies.

Resist the urge to put level 3 content at level 1. Executives need level 1; analysts need all three.


Layout principles

F-pattern reading: Eyes scan left to right, then down. Place the most important metric top-left. Put totals before breakdowns.

Proximity: Related charts should be adjacent. Group by topic, not by chart type.

White space: A crowded dashboard forces every chart to compete for attention. Remove or collapse anything that is not regularly used.

Consistent scales: Charts on the same page comparing the same metric should use the same axis scale. Inconsistent scales mislead.


Chart selection summary

Data type Recommended chart
Trend over time Line
Category comparison (≤7) Bar (vertical)
Category comparison (8+, long labels) Bar (horizontal)
Part of whole (≤4 segments) Donut
Part of whole (5+ segments) Stacked bar
Distribution Histogram or box plot
Two-variable relationship Scatter
Sequential drop-off Funnel

Colour usage

  • Use a single colour for a single series. Reserve multiple colours for multiple series.
  • Use colour to encode meaning, not decoration. Red = bad, green = good, grey = context/background.
  • Limit to 3–5 distinct colours. More colours require a legend; legends require eye movement.
  • Ensure accessibility: avoid red-green combinations that are indistinguishable to colour-blind users. Use blue-orange as the primary contrast pair.

Anti-patterns to avoid

Anti-pattern Problem
Dual y-axis charts Creates the illusion of correlation between unrelated metrics
3D charts Distorts relative sizes; never adds information
Pie charts with > 5 slices Slices < 5% are unreadable
Truncated y-axis Exaggerates differences; misleads at a glance
Too many decimal places Creates false precision; rounds to 2 significant figures
Unlabelled axes Forces the viewer to infer what is being shown
Missing time zone / period label Makes the data unverifiable

Mobile-first considerations

If the dashboard will be viewed on mobile:

  • Limit to 3 charts per page
  • Single-column layout
  • KPI cards instead of charts where possible
  • No hover interactions (use tap-to-expand)
  • Minimum 16px font size

Source: SKILL.md on GitHub

No alerts16d4 checks · Risk SAFE
  • Gen Agent Trust Hub16d

    This skill is a collection of documentation templates and design principles for dashboard creation. It contains no executable code, network operations, or external dependencies.

  • Socket16d

    No alerts

  • Snyk16d

    Risk: LOW · No issues

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at fcb0454. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 6 days ago.

Activeupdated 5 months ago

README badge

README badge for nimrodfisher/data-analytics-skills/dashboard-specification