Begin with the decision, not the chart
Suppose a small shop wants to know which product needs a stock review. An analyst first asks what a sale means, which dates count and how returns should be handled. Adding every number in a column can produce an attractive but misleading result if rows represent different things.
Data Analysis, also called Data Analytics in the approved pricing focus, concerns organising and interpreting information. In a reporting or business-intelligence setting, work can involve checking sources, agreeing definitions with colleagues and explaining an exception. Producing a chart is one deliverable; making its assumptions understandable is another.
Help a shop owner make sense of sales.
A shop owner sees five receipt rows and asks, “How much did completed orders bring in?” One receipt appears twice, one order was cancelled and one amount is missing.
Your role: You will sort the receipts before adding them. Watch the total change, then tell the owner what is known and what still needs checking.
What you’ll learn: Learn how the question and the quality of the records change a sales figure.
No account needed. Nothing is sent or saved. You can restart at any point.
Read the full guided story without the interactive controls
You can follow the same lesson here without JavaScript.
Count a purchase once
An order ID is the label for one purchase. Receipt A appears twice for the same purchase. Set aside its extra copy and watch the sum.
What happens and why: Removing the repeated A reduces the known-amount sum from 520 to 400 units. Deduplication means removing repeated records of the same thing; real data needs a reliable identifying key.
Answer the owner’s actual question
The owner asked about completed orders. Order B was cancelled. Decide whether it belongs in this calculation.
What happens and why: Setting aside cancelled B removes 80 units, leaving 320 known units for completed orders. Filtering means keeping the records relevant to a defined question.
Keep an unknown amount honest
Order C is complete, but its amount is blank. There is no evidence it was free. Compare leaving it unknown with recording zero.
What happens and why: Known completed revenue stays 320. Leaving C unknown gives an average of 160 over two known amounts; assuming zero gives 106.67 over three rows. Neither choice discovers C’s true amount. Do not claim 320 is the full total.
Give the owner a useful answer
Explain the known figure, the records you set aside and the missing amount. The owner needs the limits as well as the number.
What happens and why: The note says known completed-order revenue is 320 units from A and D. Extra A and cancelled B were excluded. C is still unknown, so complete revenue cannot yet be established.
Takeaway: Define the question, count each order once, exclude irrelevant records and disclose missing information. A clean-looking number can still be incomplete.
Where this is useful: Useful for sales reports, budgets, attendance lists or any decision based on a spreadsheet.
A fictional educational example, not a confirmed TWA assignment, assessment or learner outcome. Simplified to explore one idea; real situations need context and authorised procedures.
Optional deeper exploration: open the original technical workspace
Want to explore the technical model?
The guided story above introduces the idea. This original workspace lets you try actions in different orders. It has its own reset and does not inherit your story choices.
- Order ID
- A label used to identify one purchase. Here A appears twice for the same purchase.
- Deduplication
- Removing repeated records of the same thing.
- Filter
- Keep only records that answer the question.
- Missing value
- Information that has not been recorded; it is not automatically zero.
- Average
- A sum divided by the number of values included. Say which values you counted.
A number is only as useful as its definition.
Learning objective: Calculate revenue while making data-quality decisions explicit.
A fictional shop asks for completed-order revenue. Amounts are illustrative units, not TWA fees.
Evidence desk
- Order A
- Completed · 120
- Order A
- Completed · 120 (duplicate ID)
- Order B
- Cancelled · 80
- Order C
- Completed · amount missing
- Order D
- Completed · 200
Your workspace
Worked explanation & takeaway — available without JavaScript
Deduplicate by order ID and exclude cancelled orders: known completed revenue is 320 across two known amounts. Order C remains unknown. Treating a missing amount as zero changes an average and does not establish its real value. State exclusions before recommending a business decision.
Fictional educational model. No live system, personal assessment or confirmed TWA assignment. All actions run locally in your browser.
Change the question, change the answer
Fictional shop records; this is an arithmetic illustration, not a live business dataset.
| Record | Product | Movement | Units |
|---|---|---|---|
| S01 | Notebook | Sale | 2 |
| S02 | Pen | Sale | 1 |
| S03 | Notebook | Sale | 3 |
| R01 | Notebook | Return | 1 |
| S04 | Pen | Sale | 2 |
Count sale rows only: 2 + 1 + 3 + 2 = 8 units. Label this gross sales; returns have not been subtracted.
Read both scenarios and explanations
Gross sales
Count sale rows only: 2 + 1 + 3 + 2 = 8 units. Label this gross sales; returns have not been subtracted.
Net movement
Subtract the one-unit return from 8 sold units: net outward movement is 7 units. State the return convention explicitly.
An educational illustration using fictional situations. It is not a confirmed TWA assignment, live system, assessment or promise of a particular tool.
Build confidence with rows, questions and checks
Useful foundations include spreadsheet formulas, sorting/filtering, types of values, simple summaries and the meaning of a missing value. SQL helps express questions about structured records. Visualisation then helps communicate a comparison or pattern without hiding how it was calculated.
TWA’s existing direction names Excel, data cleaning, SQL, Power BI concepts, dashboards and communicating findings. Microsoft describes Power BI as a platform for connecting, visualising and sharing data. That is industry context; the exact software access, version, assessment and learning-plan depth must still be agreed with TWA.
A five-row example can reveal a real problem
The fictional table below contains four sales and one return. Gross sales are 2 + 1 + 3 + 2 = 8 units. Net movement subtracts the one-unit return, giving 7. Neither number is meaningful unless the label and treatment of returns are clear. The example contains no revenue or customer information.
An analyst would also check whether the records cover the whole period and whether product labels are consistent. A missing price cannot be silently treated as a zero-price sale. Document any exclusion and tell the decision-maker what evidence would change the conclusion. This reasoning transfers to larger data without making the tiny example a realistic forecasting model.
A stock-movement decision note
Use the fictional rows to explain gross sales, returns and net movement. Propose one follow-up question before anyone changes stock levels.
What you could document
- A cleaned table and simple data dictionary
- Reproducible calculations for gross and net units
- One clearly labelled chart or summary
- A decision note naming the dataset’s limitations
What makes the explanation useful?
Another person should be able to reproduce the result, understand excluded rows and distinguish observed units from predicted demand.
The learning plan determines actual activities and assessment. These deliverables are examples, not contractual curriculum outcomes.
Communicate a finding with its limits
A short analysis handover can state the question, source, calculation, finding and recommended follow-up. A stock-review recommendation should distinguish an observed movement from a claim about future demand. Correlation, incomplete coverage and data-entry problems deserve attention before a business acts.
For deeper work, compare reporting and dashboard development with Data Engineering, which builds data systems, or Data Science, which may investigate predictive models. Ask TWA to match the starting point to your spreadsheet experience and the questions you want to answer. No employment or business outcome is guaranteed.