Data Cleaning and Preparation Flashcards

In pandas, df['col'].fillna(method='ffill') does what?

Answer: Fills missing values with the previous valid value

Forward fill propagates the last valid observation forward.

Which describes 'data leakage' during preparation?

Answer: Information from the test set influences training preprocessing

Leakage occurs when test-set information leaks into training, e.g. fitting a scaler on all data.

To avoid leakage, a StandardScaler should be fit on:

Answer: The training set only, then applied to test data

Fit scalers on training data and transform test data with those parameters.

Binning a continuous 'age' column into groups like 'child', 'adult', 'senior' is called:

Answer: Discretization

Converting continuous values into discrete bins is discretization.

Which is a sign of a poorly structured ('untidy') dataset?

Answer: Multiple variables stored in a single column

Tidy data needs one variable per column; cramming several into one breaks that.

Stripping leading/trailing whitespace and fixing inconsistent capitalization in a text column is part of:

Answer: Text/string cleaning

Trimming whitespace and normalizing case are string-cleaning steps.

Median imputation is often preferred over mean imputation when the column:

Answer: Is skewed or has outliers

The median is robust to skew and outliers, unlike the mean.