mschwarzmueller/ai-agents-workflows-course/data-analytics

Data Analytics

Required reference

• Before analyzing this project's sales data, read `references/sales-data-contract.md`. It defines the data grain, field roles, sensitive columns, metric formulas, and required quality checks.

Rules

• Use only project data. Never invent numbers.
• Inspect schema, date range, and key values before calculating.
• Define each metric before reporting it. Honor project instructions when present.
• Prefer small scripts (Python/pandas or SQL) over opaque manual math.
• Keep the unit of analysis consistent (day vs month, order vs customer, etc.).
• Lead with the answer; keep supporting detail brief.

Workflow

1. Restate the question: metric, filters, time range, and breakdown.
2. Read `references/sales-data-contract.md` before using sales.csv; apply its definitions and checks.
3. Profile the relevant files and columns: types, nulls, categories, and date bounds.
4. Apply explicit metric logic and filters; run the calculation.
5. Sanity-check row counts, date bounds, excluded statuses, currencies, duplicates, and outliers.
6. Report: direct answer → short breakdown → one-line methodology note containing source, filters, definition, and caveats.

When reporting

If the user wants a polished HTML report or dashboard page, hand off to create-report with the final numbers and definitions. Do not style the report in this skill.