Skip to content

hoavdc/CodexKit / skills/codexkit-data-story-builder

codexkit-data-story-builder

Unverified

Turn business data, KPI movement, or experiment results into a clear narrative using the what-so what-now what structure, audience calibration, and action-oriented insight. Use when leaders need a data-backed brief, dashboard storyline, or metric interpretation. Do not use for raw statistical modeling with no communication deliverable.

Data Story Builder

Purpose

Make analytics usable by giving the numbers a decision-oriented narrative.

When to use

  • A dashboard or KPI movement needs interpretation.
  • Experiment results or trend shifts must be explained to leadership.
  • A team needs a data-backed narrative, not a raw chart dump.

When not to use

  • The task is purely technical modeling with no stakeholder communication output.
  • The available data is too weak to support any claims and the user refuses caveats.

Inputs

  • business question and target audience
  • data points, charts, or KPI movement
  • baseline, target, or expected benchmark
  • context events that may explain the movement

Procedure

  1. Start from the business question, not the chart.
  2. Separate signal, uncertainty, and noise.
  3. Structure the story as what happened, why it matters, and what to do next.
  4. Translate numbers into plain-language implications for the chosen audience.
  5. Recommend the next decision, experiment, or investigation.
  6. State confidence limits and missing data.

Output

  • headline insight
  • what changed
  • why it matters
  • likely drivers or interpretations
  • recommended next actions
  • caveats and confidence notes

Definition of done

  • The audience can act on the analysis.
  • The narrative separates evidence from interpretation.
  • Caveats are present where the data is weak.

Examples

  • "Turn this KPI dashboard into a narrative for the monthly business review."
  • "Explain these A/B test results for a non-technical leadership team."

Quality Criteria

  • The story starts from a business question, not from chart narration.
  • Evidence, interpretation, and recommendation are clearly separated.
  • Every claim is supported by a data point, comparison, or stated assumption.
  • The "so what" explains business consequence, not just metric movement.
  • Caveats and confidence limits are included when data is incomplete or noisy.

Verification (4C)

CheckQuestion
CorrectnessDo the numbers, comparisons, and causal language match the underlying data?
CompletenessDoes the story include what changed, why it matters, likely drivers, actions, and caveats?
Context-fitIs the narrative useful for the audience's actual decision or operating review?
ConsequenceWhat wrong action might a stakeholder take if the story overstates certainty?

Edge Cases

  • Correlation mistaken for causation — Use causal language only when the design supports it; otherwise state "may be associated with."
  • Metric definition changed — Split the story before and after the definition change.
  • Executive audience with little time — Lead with the decision implication, then supporting evidence.
  • Weak or missing baseline — Mark confidence as low and recommend the next analysis step.

Changelog

  • v1.1.0 — Added data-story-specific quality gates and consequence checks.
  • v1.0.0 — Initial release
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 464kUpdated Claude CodeCodexVerifiedInstall
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 464kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 318kUpdated Claude CodeCodexUnverifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 318kUpdated Claude CodeCodexUnverifiedInstall