The point of analytics isn't to produce charts — it's to change decisions. Dashboards, reports, metric definitions, data visualizations, and statistical analysis all need to answer the question: "so what?" These skills cover the last mile of data work: turning numbers into narratives, building dashboards that people actually use, and defining metrics that measure what matters instead of what's easy to count.
Key takeaway
ClaudeVault's analytics and reporting skills give Claude Code structured workflows for the last mile of data work — dashboard design that limits screens to five to nine metrics to avoid the 40 percent engagement drop from cognitive overload, metric definition frameworks that prevent teams from measuring what is easy instead of what matters, data visualization that matches chart type to data relationship, statistical analysis with significance testing and confidence intervals, report generation for recurring business reviews, and data storytelling that turns numbers into narratives that change decisions.
When dashboards have 30 metrics per screen and stakeholders ignore them because finding the signal requires more effort than asking an analyst directly
When two teams define the same metric differently and reports conflict because there is no single source of truth for metric definitions
When data visualizations use the wrong chart type — pie charts for time series, bar charts for correlations — and misrepresent the underlying relationships
When analytics reports describe what happened last quarter but never explain why it happened or what the team should do about it
A Claude Code analytics workflow builds from metric definition through visualization to narrative, ensuring every dashboard and report answers the question that matters: so what?
Start with the metrics definition advisor. Claude defines each metric with its formula, data source, granularity, owner, and the question it answers. The output becomes the single source of truth — when two teams disagree on a number, the metric definition resolves the conflict before the meeting starts.
The dashboard designer builds screens for specific audiences — operational dashboards for daily decisions with real-time data, strategic dashboards for quarterly reviews with trend lines and targets. Claude limits each screen to five to nine metrics, groups related indicators, and designs the layout so the most important signal is visible without scrolling.
Use the data visualizer to match chart type to data relationship. Claude selects line charts for trends, bar charts for comparisons, scatter plots for correlations, and heatmaps for density — never pie charts for more than five categories, never dual-axis charts that obscure the scale. Each visualization has a title that states the insight, not the data series.
The statistical analysis advisor applies the right test to the question. Claude selects t-tests for group comparisons, chi-squared for categorical relationships, regression for predictive modeling, and calculates confidence intervals so the analysis reports uncertainty alongside the estimate.
Finally, the data storytelling advisor transforms metrics and visualizations into a narrative that changes decisions. Claude structures the story as context, tension, and resolution — what the audience expected, what actually happened, and what the team should do about it — so the report drives action instead of just documenting the past.
Outcome
Metrics defined with shared formulas and owners, dashboards designed for specific audiences with cognitive load limits, visualizations matched to data relationships, statistical analysis with significance testing, and a narrative layer that connects the numbers to decisions — analytics that answers so what.
| Skill | Best for | Complexity | Primary use case |
|---|---|---|---|
| Metrics Definition Advisor | Shared metric standards | Intermediate | Metric formulas, data sources, owners, and single-source-of-truth definitions |
| Dashboard Designer | Decision-focused screen layouts | Intermediate | Operational and strategic dashboards with 5-9 metrics per screen |
| Data Visualizer | Chart type selection and design | Beginner | Matching visualization type to data relationship and insight communication |
| Statistical Analysis Advisor | Significance testing and modeling | Advanced | T-tests, chi-squared, regression, and confidence interval calculation |
| Report Builder | Recurring business review documents | Beginner | Structured reports for weekly, monthly, and quarterly business reviews |
| Data Storytelling Advisor | Narrative-driven analytics communication | Intermediate | Context-tension-resolution narratives that drive decisions from data |
Designs analytics dashboards with KPI selection, chart type decisions, layout hierarchy, and drill-down paths. Use when building dashboards that support specific business decisions. Dashboard layout, refresh strategy, BI design.
**Audience:** [Who] **Primary question:** [What decision] **Refresh cadence:** [How often] | KPI | Definition | Target | Comparison | Chart Type | [Structured description of rows/columns] [Chart
Structures data presentations with narrative frameworks that drive decisions. Use when presenting analytical findings to stakeholders. SCRA framework, audience calibration, chart annotation, actionable recommendations.
Turn analytical findings into narratives that drive action — leading with insight, framing for the audience, annotating charts to be self-explanatory, and ending with specific recommendations.
Recommends chart types and generates visualization code with data storytelling principles. Use when presenting data visually in reports, dashboards, or presentations. Chart selection, matplotlib, plotly, recharts.
Recommend the right chart types for given data, generate visualization code, and apply storytelling principles that make data communicate clearly — choosing visualizations that reveal patterns, not ob
Defines business metrics with mathematical precision to eliminate conflicting calculations across teams. Use when multiple teams report different numbers for the same metric. Metric specification, edge cases, source of truth, dimensional breakdowns.
Define metrics with mathematical precision — specifying exact numerator, denominator, time window, filters, and edge cases so every dashboard, report, and query produces the same number.
Structures analytical data into professional reports with executive summaries, trend analysis, and actionable recommendations. Use when turning raw data or query results into decision-driving reports. Report structure, metrics tables, insight framing.
Take raw data, analysis results, or query outputs and structure them into clear analytical reports — with executive summaries, formatted tables, trend analysis, and actionable recommendations that tur
Recommends statistical tests, calculates sample sizes, and interprets results with business context. Use when running experiments, A/B tests, or analyzing whether observed differences are real. Test selection, power analysis, significance interpretation.
Recommend the right statistical test, calculate required sample sizes, interpret results correctly, and flag common misinterpretations that lead to bad decisions — translating statistics into business
Five to nine metrics per screen. Research shows that exceeding working memory capacity causes a 40 percent engagement drop — users stop looking at dashboards that feel overwhelming. The dashboard designer groups related metrics, prioritizes the most important signal at the top, and uses a drill-down structure for detail instead of cramming everything onto one page.
A metric definition framework documents each metric's formula, data source, granularity, owner, and the business question it answers. It matters because without one, two teams will define revenue or churn differently and produce conflicting reports. The metrics definition advisor creates the shared definitions that prevent these conflicts.
Looker embeds a semantic modeling layer and excels at governed enterprise analytics. Tableau has the strongest visualization design capabilities. Power BI integrates with the Microsoft ecosystem at the lowest per-user cost. Metabase is open-source and ideal for startups that want self-hosted BI. The dashboard designer skill generates dashboard specifications that work with any of these tools.
Data storytelling structures analytics as a narrative — context (what the audience expected), tension (what actually happened and why it differs), and resolution (what the team should do about it). It bridges the gap between producing charts and changing decisions. The data storytelling advisor transforms dashboards and reports into narratives that drive action.
Operational dashboards serve daily decisions with real-time or near-real-time data, fewer metrics, and alert thresholds. Strategic dashboards serve quarterly reviews with trend lines, targets, and variance commentary. Executive dashboards show the three to five numbers that matter most with drill-down access to detail. The dashboard designer builds each type for its specific audience and decision cadence.
A metrics layer is a shared computation layer — tools like dbt Metrics, Cube.js, or PowerMetrics — that defines metric formulas once and exposes them consistently to every BI tool and consumer. You need one when different dashboards calculate the same metric differently, or when business logic lives in dashboard-specific SQL instead of a central definition.