From business questions to an analysis plan.
Turn an unclear stakeholder request into a neutral question, a metric contract and a reproducible analysis plan.
A better brief.
A defensible plan.
Translate a leading stakeholder request into a neutral question. Define the population, measures, quality checks and evidence needed before recommending an operational change. A headline rate can improve while rates within each channel stay unchanged.
Before you start: Data Analyst Foundations or equivalent comfort with counts, percentages and basic data tables. No Python, SQL or paid software is required.
Allow about 4 hours and 15 minutes, including exercises. Work with a calculator, text editor or local spreadsheet. Every lesson includes a detailed explanation, practical task, worked solution and knowledge check.
The extract is prepared source data with meaningful blanks, not a cleaning challenge. You will plan analysis, verify a small arithmetic preview and distinguish descriptive evidence from causal claims.
Five lessons. One clear plan.
Turn a vague request into a decision brief
Separate the decision, analytical question and requested output.
Define scope, grain and fair comparison windows
Define the unit, cohort boundaries and observation cutoff.
Specify metrics, denominators and guardrails
Write a numerator, denominator and missing-data rule for each measure.
Request the evidence and plan quality checks
Map questions to fields and identify unavailable evidence.
Write and review a defensible analysis plan
Sequence methods, validation and deliverables in one plan.
A real practice task.
A manageable dataset.
Read the case brief and dictionary, then draft your own plan and metric contract. Use the separate worked solutions after attempting the exercises. The complete pack contains twelve offline files, including a styled worked plan with relative evidence links.
Snapshot: 22 June 2026 00:00 UTC. The cohorts are 1–7 and 8–14 June. Missing ratings remain missing. The same fictional case connects the Applied modules and is separate from Northstar Shop. No real company outcome or customer data is represented.
Could another analyst follow your plan?
Use the ten-point rubric to check neutrality, scope, metrics, methods and interpretation. Aim for eight points with no zero in scope or metric correctness and no unsupported causal claim. This is a self-review guide, not an accredited assessment or operational rollout approval.
Begin Module 01 →Return to Applied Data Analysis ↗
Continue to Module 02: Exploring data and finding useful patterns →
Explore practical lessons, worked examples and projects. Follow a published path, practise the method and build work you can explain. Keep your work locally; no upload is required.