Practice — Mild · Solution
The five diseases (step 2's list)
| Disease | Verified extent | Tool |
|---|---|---|
| Exact duplicate rows | 14 (514 → 500) | Remove Duplicates, all columns |
| Trailing spaces in Name | 49 names | Format → Trim |
| Text "N/A" in Segment | 22 cells | Replace Values → true null |
| Spelled-out states | 29 rows, 4 values (30 → 26 distinct) | Replace Values ×4 |
| Mixed date formats in Joined | 178 / 171 / 165 across three formats | Lesson's §5 — typed in A02 |
All figures from the raw-file profile — and your checkpoints should have matched them exactly: 514→500 rows, 22 empty Segments, 26 distinct States, card reading 500.
Reflection answers
- A defensible ranking: the duplicates first (every count and total inflated, invisibly), the spelled-out states second (every state-level chart splits its bars), the fake N/As third (filters and slicers lie), trailing spaces fourth (joins and counts split on lookup), and the date formats last only because the column was not yet used — the moment it is, it jumps to first. Reasonable people can reorder; the point is being able to argue consequences, not tidiness.
- Zero. That is the whole answer, and the whole tool: the recipe replays on refresh. If your instinct said "all seven," the mental model to install is: you did not clean a file, you wrote a program.