DA0-002 Data Acquisition and Preparation Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✓

  dropna()

  Why this is correct

  The `dropna()` method removes rows or columns containing null values, directly satisfying the requirement to handle missing entries in the 'age' and 'income' columns. It suits scenarios where incomplete records should be excluded entirely, though it risks discarding otherwise valid data when nulls are sparse across the DataFrame.

• ✗

  pivot_table()

  Why it's wrong here

  pivot_table() reshapes data into a cross-tabulated summary; it does not impute or remove nulls. It tempts because pivoting often precedes reporting and can expose gaps visually, so it would be the right method when aggregating values by category, not when handling missing entries.

• ✗

  merge()

  Why it's wrong here

  merge() joins two DataFrames on shared keys; it neither detects nor fills nulls in 'age' and 'income'. It tempts because analysts routinely merge reference tables to enrich data before cleaning, so it would be correct when combining datasets, not when imputing or dropping missing values.

• ✓

  apply() with a custom function

  Why this is correct

  apply() with a custom function lets you define bespoke imputation logic per column, such as filling 'age' with the median and 'income' with a group-wise mean, handling the missing values precisely as the scenario requires.

• ✓

  fillna()

  Why this is correct

  fillna() replaces missing entries with a specified value or a computed statistic, such as the mean age or median income, satisfying the requirement to handle NaN values without discarding rows. Unlike dropna(), it preserves record count, which suits the analyst's need to retain data while imputing gaps in the 'age' and 'income' columns.