GOkul069

Data Storytelling — Data skill for Claude Code

Data community

Claude skill that turns analyses and Tableau/Power BI dashboards into narratives for execs, decks, Medium/LinkedIn, and docs.

How to install Data Storytelling

This entry records only its repository, not the path inside it, so there is no exact command to give. Open GOkul069/data-storytelling and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Data Storytelling does

Claude skill that turns analyses and Tableau/Power BI dashboards into narratives for execs, decks, Medium/LinkedIn, and docs.

Alternatives in Data

  • Osint Employees — Gather employee names + email patterns for password spray prep 4.5k ★
  • L Docs Architecture — Document cross-cutting Lowdefy architecture topics (build pipeline, request lifecycle, etc 3k ★
  • Portaljs Add Resource — Add another file (resource) to an EXISTING dataset in a PortalJS portal — a data dictionary, methodology, or a 2.3k ★

README

Data Storytelling

A Claude skill that turns an analysis or a Tableau / Power BI dashboard into a story an audience will act on. It writes for one of four audiences:

Format Who it's for What you get
Executive summary Leadership, VPs, board A memo of 250 words or less: the bottom line, what happened, why, the recommendation and watch-outs
PowerPoint deck Presentations, QBRs A 6–10 slide storyline whose titles tell the story on their own, with chart specs and speaker notes (or a real .pptx when a slides tool is available)
Medium / LinkedIn post Public audiences A LinkedIn post (150–300 words) or a Medium article (800–1,500 words), with confidential numbers checked and anonymised
Python / technical docs Analysts and engineers A README-style write-up: data, metric definitions, method, how to reproduce, plus pandas versions of DAX or Tableau calcs

Why it exists

Dashboards show numbers; they don't say what the numbers mean. Stories written straight from a dashboard tend to go wrong in a few predictable ways:

  • They give a tour of the dashboard ("the top-left chart shows…") instead of leading with the finding.
  • They quote numbers without saying what they're compared to.
  • They present a hunch about the cause as if it were fact.
  • They mistake calendar effects for real changes (a 28-day February looks like a crash in monthly totals).

The skill makes Claude build a fact sheet first and write the story second, then runs a quality check before handing the result back.

How it works

  1. Pick the format. It uses the one you name. If you don't name one, it infers it from cues ("board" → exec summary, "slides" → deck, "LinkedIn" → post, "README" → docs), and falls back to an executive summary.
  2. Extract the facts. It reads screenshots, Tableau .twb/.twbx XML, Power BI measures and slicers, CSVs, notebooks or pasted metrics. For each metric it records the value, the comparison, the chan