Data Cleaning Techniques

Name:

Date:

📚 Part 1: Multiple Choice & Matching

1. What is the best short definition of data cleaning?

Removing all data so the set is smaller

Finding and correcting errors or inconsistencies so data is ready for analysis

Creating new variables without checking original values

2. Which of the following is an example of a data accuracy problem? (Choose one)

Consistent use of the same units across a column

Typos in text entries (e.g., "Auckland" and "Auckland")

All numeric values recorded to the same precision

3. Check all methods that can be sensible ways to handle missing numeric values in a dataset:

Remove the entire row with the missing value

Impute a value (e.g., mean or median) with justification

Replace missing values with an arbitrary number like 0 without justification

Keep missing values and record why they are missing

4. Match the term (left) with the correct description (right). Draw lines between matching items.

A. Imputation

B. Recategorisation

C. Outlier

D. Duplicate row

1. Converting '17', '17 years' to a single age category

2. A data point that is very different from others and may need checking

3. Filling a missing value using a calculated estimate (e.g., mean)

4. Two identical survey responses recorded twice

✏️ Part 2: Short Answers & Practical Tasks

5. Look at these survey entries for Age: 16, 17, 15, "sixteen", 18. List the inaccuracies or inconsistencies you see and write two steps to clean them.

6. Convert the following distance values into kilometres and write the cleaned list (one value per item): 5 km, 5000 m, 3.2 km, 3200 m.

7. You find a column labelled "Salary" with one entry recorded as "$50,000" and another as "50000". Explain a simple cleaning step to make these consistent.

8. A dataset has a missing test score for one student. Briefly justify whether you would remove the row, impute a value, or leave it blank. Include one reason related to how your choice affects later analysis.

9. State one ethical consideration to remember when cleaning real personal data (one brief sentence).