Pilot · Lab P06 · 45 minutes
Cleaning journey records
A fictional class logged six journey records, with a duplicate and a missing value among them. Decide which rows to include, log each cleaning decision, calculate the mean journey time and the bus mean, and write a claim that is qualified by what the data can and cannot show.
Predict first
Write your predictions before you use the tool.
- Look at the six rows. Which ones would you leave out of a mean journey time, and why?
- Predict the mean journey time after keeping one row per trip and excluding the missing value.
- Predict what happens to the mean if the missing minutes are treated as 0.
Practice tool
The journeys tool applies two cleaning choices to the six fixed fictional rows: keep one row per trip ID or keep duplicates, and exclude missing minutes or treat them as 0. It then shows each row’s decision with a reason, plus the count, mean and bus mean. It contains no real personal data and does not decide which claim is true.
Journey records · fixed six-row dataset
| Row | Trip | Minutes | Mode | Decision |
|---|---|---|---|---|
| 1 | A | 10 | walk | Included: Valid |
| 2 | B | 20 | bus | Included: Valid |
| 3 | C | 20 | bus | Included: Valid |
| 4 | C | 20 | bus | Included: Valid |
| 5 | D | missing | walk | Included: Missing treated as 0 (unsafe) |
| 6 | E | 50 | bus | Included: Valid |
Rows used: 6 · mean minutes: 20 · bus mean: 27.5
Starter
trip, minutes, modeA, 10, walkB, 20, busC, 20, busC, 20, busD, (missing), walkE, 50, busrow, included?, reason1, , …Worked case
| row | trip | minutes | decision | reason |
|---|---|---|---|---|
| 1 | A | 10 | include | valid |
| 2 | B | 20 | include | valid |
| 3 | C | 20 | include | valid |
| 4 | C | 20 | exclude | duplicate of trip C |
| 5 | D | missing | exclude | minutes missing |
| 6 | E | 50 | include | valid |
Four valid rows: (10 + 20 + 20 + 50) ÷ 4 = 25 minutes. Bus rows B, C, E: (20 + 20 + 50) ÷ 3 = 30 minutes.
Missing is not 0. Treating D as 0 minutes keeps five rows and pulls the mean down to 20, as if someone arrived instantly. Keeping the duplicate counts trip C twice. Each choice changes the result, so the cleaning log is part of the evidence.
A qualified claim: "In these four recorded trips, the bus trips averaged 30 minutes and the one walk averaged 10." It does not say buses cause longer journeys: distance was not recorded, and four trips cannot represent the whole school.
Build
- Fill in the cleaning log for all six rows before using the tool.
- Set the tool to one row per trip and exclude missing; compare its decisions with your log.
- Record the count, mean and bus mean.
- Try the other three combinations and record how each result changes.
- Write one claim the data supports and one limitation that keeps it from being a general conclusion.
Tests
| Input | Expected | Why this test |
|---|---|---|
| One per trip, exclude missing | 4 rows; mean 25; bus mean 30 | The intended cleaning. |
| Keep duplicates, exclude missing | 5 rows; mean 24; bus mean 27.5 | Trip C is counted twice. |
| One per trip, missing as 0 | 5 rows; mean 20; bus mean 30 | An invented 0 drags the overall mean down; the walk row is not a bus row. |
| Keep duplicates, missing as 0 | 6 rows; mean 20; bus mean 27.5 | Both errors at once. |
Explain and reflect
Explain
- Explain, with the numbers, why treating missing minutes as 0 is not a neutral choice.
- Explain why your claim describes an association in these trips and not a cause.
Reflect
- What extra column would most improve the claim about bus and walking times?
- How many more trips, collected how, would you want before generalising to the school?
What this lab does not do
- The six rows are fixed and fictional; you cannot add or edit rows in the tool.
- The tool offers only two cleaning choices; real data often needs more.
- Nothing is saved.