Week 1 — Questions, Data & Spreadsheets · ~50 min
The analyst's question
Free previewWhat you'll learn this week
- Turn a vague request ("look at our sales") into a precise analytical question
- Name the stakeholder, the decision, the population, the period, the row grain and the metric
- Recognise the six kinds of field you will meet in every dataset
- Inspect where data came from, what it covers, and — just as important — what it does not contain
- Navigate a spreadsheet safely: sheets, tables, ranges, the formula bar, and formats vs stored values
- Write formulas with relative and absolute references and check them after copying
Analysis starts before the data
A shop owner in Ibadan says: "Can you look at our sales and tell me what's going on?"
That is not a question an analyst can answer. It is an invitation to have a conversation. A beginner opens the spreadsheet and starts making charts. A professional first asks:
- Who will use the answer? (the stakeholder)
- What will they do differently depending on the answer? (the decision)
- Which things are we counting — all orders, only completed ones, only one branch? (the population)
- When — last month, last quarter, since we opened? (the period)
- What does one row represent? (the grain)
- What number will summarise the answer — revenue, number of orders, average order value? (the metric)
Rewritten, the request becomes:
The shop owner (stakeholder) wants to decide which product category to restock first next month (decision). Using all order lines (population) from January to March 2026 (period), where each row is one order line (grain), which category produced the most revenue (metric)?
That sentence is long, but every part of it changes the analysis. Change "revenue" to "number of orders" and you might get a different winner.
Separate the question from the answer you want
Stakeholders often arrive with a preferred answer: "Show me that Books are our best category." Your job is to write the question so that any answer is possible:
| Leading (bad) | Neutral (good) |
|---|---|
| Prove that Books outsell Stationery | Which category had higher revenue, Jan–Mar 2026? |
| Show that the North region is struggling | How does revenue compare across regions, Jan–Mar 2026? |
| Why are customers leaving? | How many customers placed no order in the period? |
The last example matters: "why are customers leaving" assumes they are leaving and that the data can explain why. Most sales data can tell you what happened. It rarely tells you why.
The practice dataset
All 12 weeks use the same small, fictional dataset. It is small on purpose: you can check every answer by hand. Monetary values are in an arbitrary currency — think of it as ₦ thousands if that helps.
customers.csv
customer_id,customer_name,region
1,Amina,North
2,Bola,South
3,Chidi,North
4,Dara,West
5,Efe,South
orders.csv — each row is one order line. Revenue = quantity × unit_price.
order_id,customer_id,order_date,category,region,quantity,unit_price
101,1,2026-01-05,Books,North,2,10
102,2,2026-01-12,Stationery,South,5,3
103,1,2026-02-02,Books,North,1,15
104,3,2026-02-10,Stationery,North,4,5
105,4,2026-02-18,Books,West,3,10
106,2,2026-03-03,Books,South,2,15
107,3,2026-03-15,Stationery,North,2,5
108,4,2026-03-21,Books,West,1,20
Memorise these checks. You will use them every week to catch your own mistakes:
- orders has 8 rows
- total revenue is 160
- customer 5 (Efe) has no orders
- monthly revenue: January 35, February 65, March 60
If your Power BI report in Week 7 says total revenue is 320, you will know immediately that something has doubled — and you will know it because you checked by hand in Week 1.
Rows, observations, and grouped results
Three words people mix up:
- A row is a physical line in the table.
- An observation is the real-world thing a row records. In orders.csv, one row = one order line (one product category bought in one order).
- A grouped result is a summary row you create — e.g. "North: 65". It represents many observations, not one.
When someone says "North has 65", ask: 65 what? Revenue from 4 order lines. That "North" row is not a customer and not an order. Confusing grouped results with observations is the root of half the mistakes in this course.
Self-check
- Rewrite "How is the business doing?" as a precise question with all six parts.
- In orders.csv, what is the grain? In customers.csv?
- Explain the difference between a row, an observation, and a grouped result, using the North region as your example.
Sign in to track your progress.