Data Analysis

This is a structured workflow for turning messy datasets into decision-ready findings without skipping the boring parts like data quality checks and caveat documentation. It pushes you to frame the analysis question first, then pick the cheapest lane that works: spreadsheet triage, SQL slicing, notebook stats, or just writing up what you already know. The instructions cover change explanation, segment comparison, funnels, and telemetry with explicit checklists for each. What stands out is the insistence on separating observation from interpretation and actually documenting trust level before promising conclusions. If you've ever been handed a CSV export with vague instructions or needed to explain an A/B test without overselling the result, this gives you guardrails for the whole process.