Data Cleaning Techniques Quiz

5 questions Pass: 70% +25 pts (50 with Pro)

Quiz covering Feature Engineering

What you'll be asked

  1. When handling missing numerical data, which of the following is the most common 'easy' approach to fill gaps without significantly altering the dataset's distribution?
  2. You are preparing a dataset where a feature represents 'Annual Income' and contains several extreme outliers. Which scaling technique would be most appropriate to ensure these outliers do not disproportionately influence your model?
  3. When performing One-Hot Encoding on a categorical feature with high cardinality (e.g., 'Zip Code'), what is the primary risk?
  4. Why is it important to perform data splitting (train/test split) BEFORE applying data normalization or imputation techniques?
  5. You are dealing with a dataset where a feature has a highly skewed distribution (long tail). You decide to apply a log transformation. What is the mathematical requirement for this operation to be valid?

Answers and explanations come after you submit.