Practice Exercises — Introduction to Data Analytics¶

These exercises are conceptual — no tools required. They build the single most important analyst habit: framing a business question before touching data. Work through them with a notepad.

How to Use¶

Write your answers down. The discipline of articulating the analytics type, the decision, and the stakeholder is exactly what separates analysts who get listened to from those who produce ignored reports.

Level 1 — Warm-Up: Classify the Analytics Type¶

For each request, identify whether it is descriptive, diagnostic, predictive, or prescriptive.

"How many orders did we ship last month?"
"Why did our conversion rate drop in March?"
"Which customers are most likely to churn next quarter?"
"What discount should we offer to maximise profit?"
"Show me revenue by region for the last year."
"What caused the spike in support tickets last week?"
"Forecast next quarter's headcount needs."
"Which marketing channel should we invest the next £50k in?"

Show answers

Descriptive (what happened)
Diagnostic (why did it happen)
Predictive (what will happen)
Prescriptive (what should we do)
Descriptive
Diagnostic
Predictive
Prescriptive

Level 2 — Frame the Decision¶

For each scenario, write down: (a) the business question, (b) who the stakeholder is, (c) what decision the analysis enables, and (d) the analytics type.

Scenario A: The marketing team noticed website traffic is up 20% but sales are flat.

Scenario B: The CFO wants to know if the company will hit its annual revenue target.

Scenario C: Customer support is overwhelmed and wants to know which product generates the most complaints.

Show example answers

Scenario A - Question: Why has increased traffic not converted to sales? - Stakeholder: Head of Marketing - Decision: Where to focus optimisation — traffic quality, landing page, or checkout - Type: Diagnostic

Scenario B - Question: Based on current run-rate, will we hit the annual target? - Stakeholder: CFO - Decision: Whether to adjust spending or sales targets mid-year - Type: Predictive (forecasting)

Scenario C - Question: Which product drives the most support complaints? - Stakeholder: Head of Support / Product - Decision: Which product to prioritise for a quality fix - Type: Descriptive (leading into diagnostic)

Level 3 — Stretch: Spot the Anti-Pattern¶

Each of these describes an analytics failure. Identify what went wrong and how you'd fix it.

An analyst spends two weeks building a beautiful churn dashboard. Nobody uses it. When asked, the sales team says "we already knew customers were leaving — we needed to know which ones to call."
A report shows that "ice cream sales correlate with drowning deaths" and recommends regulating ice cream.
An analyst presents a 40-slide deck with every chart they made. The executives leave without making a decision.
A dashboard shows "Revenue: £48,250" with no comparison, target, or trend. The CEO asks "is that good?"

Show answers

No decision framing. The analyst built output, not a decision tool. Fix: ask "what decision does this enable?" first — the team needed an actionable at-risk list, not a trend dashboard.
Correlation mistaken for causation. Both are driven by hot weather (a confounder). Fix: never recommend action from correlation alone; look for the confounding variable.
Data dump, no narrative. The analyst showed their work instead of the answer. Fix: lead with the insight and recommendation; put detail in an appendix.
No context. A number without comparison is meaningless. Fix: always show vs target, vs previous period, or as a trend.

Reflection¶

After these exercises, you should be able to answer, for any request that lands on your desk:

What type of analytics is this?
What decision will it enable, and for whom?
What's the simplest analysis that answers the question?
How will I present it so the decision actually gets made?

If you can answer these before opening a tool, you're already thinking like a senior analyst.

Previous: 05-real-world-case-studies | Next: 07-interview-questions