# Data Analyst

> Use when performing data analysis, generating reports, building dashboards, writing SQL queries for business insights, or interpreting data trends. Trigger phrases: "analyze data", "data report", "SQL query for insights", "business metrics", "KPI dashboard", "Excel/Sheets analysis", "Tableau", "Looker", "Power BI", "data visualization", "cohort analysis", "funnel analysis", "A/B test results", "pivot table", "ad-hoc analysis".

- Skill: `barastrong/data-analyst` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add barastrong/data-analyst`
- Raw SKILL.md: https://api.skillmd.com/api/skills/barastrong/data-analyst/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- Author: barastrong (https://skillmd.com/u/barastrong)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/barastrong/data-analyst

---


# Data Analyst

## Role Summary

A Data Analyst extracts, transforms, and interprets data to answer business
questions, drive decisions, and communicate insights through reports,
dashboards, and presentations. The role bridges raw data and business value.

## Core Responsibilities

- Write SQL queries to extract and aggregate data from data warehouses
- Build dashboards and reports in BI tools (Tableau, Power BI, Looker, Metabase)
- Define and track KPIs and business metrics
- Conduct ad-hoc analysis to answer business questions
- Perform statistical analysis (trends, correlations, cohorts, funnels)
- Validate data quality and flag anomalies
- Present findings to stakeholders with clear narrative
- Collaborate with Data Engineers on data model improvements

## Standard Workflow

1. **Understand the Question** — clarify business question, success metric,
   deadline, and audience.
2. **Data Discovery** — identify relevant tables/sources; check data
   quality and completeness.
3. **Query & Extract** — write optimized SQL; use CTEs for readability.
4. **Transform & Aggregate** — clean, pivot, and aggregate in SQL or Python.
5. **Analyze** — identify trends, outliers, segments, correlations.
6. **Visualize** — choose appropriate chart type per data type and audience.
7. **Narrate & Present** — insight-first structure: finding → evidence →
   recommendation.

## Technology Stack

| Layer | Tools |
|-------|-------|
| Query | SQL (PostgreSQL, BigQuery, Snowflake, Redshift, dbt) |
| BI/Dashboards | Tableau, Power BI, Looker, Metabase, Superset, Grafana |
| Programming | Python (pandas, matplotlib, seaborn, plotly), R |
| Spreadsheets | Excel, Google Sheets |
| Statistics | scipy, statsmodels, basic R |
| Collaboration | Jupyter Notebook, Hex, Observable |

## Best Practices

- Define the question precisely before writing any query.
- Cite source tables and last-refresh time in every report.
- Use CTEs over subqueries for complex SQL — readability matters.
- Segment before aggregating — averages can hide important patterns.
- Always include sample size / confidence interval with percentages.
- Insight-first: lead with the answer, support with the data.
- Reproducibility: save queries and notebooks, not just screenshots.
- Data quality check: NULL counts, duplicate checks, range validation.

## Anti-Patterns to Avoid

- Cherry-picking data to confirm a pre-existing conclusion.
- Showing raw numbers without context (vs. prior period, target, baseline).
- Precision theater: "42.37%" when "~42%" suffices.
- Missing axis labels, legends, or source attribution on charts.
- SELECT * in production queries.
- Overloading one dashboard with 20+ charts.

## References

- `references/sql-patterns.md` — common SQL patterns (cohort, funnel, retention)
- `references/chart-selection-guide.md` — which chart for which data type
- `references/statistical-significance.md` — A/B testing and significance basics

## Expected Output Format

1. SQL query (clean, with comments for non-obvious logic)
2. Dashboard or report with source and refresh date noted
3. Written narrative: key findings + recommendations
4. Data quality notes (known limitations or caveats)