Craft Compelling Data Narratives

This skill turns raw data into a narrative built on a setup-conflict-resolution structure and a six-beat arc: a hook that grabs attention with a surprising insight, context establishing the baseline, rising action through data points, the key insight as climax, recommendations as resolution, and a call to action. It rests on three pillars - data as evidence (numbers, trends, comparisons), narrative as meaning (context, causation, implications), and visuals as clarity (charts, diagrams, highlights) - and provides three reusable frameworks: a Problem-Solution story (hook, context, problem, insight, solution, expected impact, call to action, as in a churn-analysis example), a Trend story (where we started, what changed, the transformation shown as a before/after table, the key insight, going forward), and a Comparison story (the question, side-by-side factors, a weighted scoring matrix, the recommendation, risk mitigation, as in a market-expansion example). It gives transition-phrase vocabulary for building the narrative, introducing insights, and moving to action, and a best-practices Do's/Don'ts list: lead with the "so what," use the rule of three, show don't tell, make it personal, and end with action - versus data-dumping, burying the insight, using jargon, leading with methodology, or dropping the narrative.

Use it when presenting analytics to executives, building quarterly business reviews or investor presentations, writing data-driven reports, communicating insights to non-technical audiences, or making a data-backed recommendation. It does not apply outside data storytelling - a different domain or tool is out of scope.

Inputs: raw data, metrics, and the business question or recommendation the story needs to support. Outputs: a chosen story framework filled in with the specific data, a visualization built with one of three techniques - progressive reveal (adding chart layers slide by slide), contrast-and-compare (before/after or side-by-side panels), or annotation and highlight (a matplotlib chart with event annotations, a highlighted growth region, and a target threshold line) - and a presentation assembled from one of three templates: an executive summary slide (key insight, supporting data, and implications side by side), a seven-slide data story flow (headline, context, discovery, deep dive, recommendation, impact, ask), or a one-page dashboard story (headline, key metrics at a glance, what's working, what needs attention, root cause, recommendation, next month's focus). Headlines follow the formula [specific number] + [business impact] + [actionable context], and uncertainty is stated explicitly with confidence levels or ranges rather than false precision.

Matplotlib for annotated, highlighted charts (ax.annotate, ax.axvspan, ax.axhline), and external references on the discipline: Cole Nussbaumer's "Storytelling with Data," Barbara Minto's "The Pyramid Principle," and Nancy Duarte's "Resonate."

ax.annotate(
    'Product Launch\n+32% spike',
    xy=(launch_date, launch_revenue),
    xytext=(launch_date, launch_revenue * 1.2),
    arrowprops=dict(arrowstyle='->', color='#E63946'),
)

Analysts, product managers, and founders who need to turn a data analysis into a narrative that drives an executive, investor, or non-technical decision rather than a dump of charts and numbers.