Data Science Data Cleaning and Preparation 5

Question 1

Combining two DataFrames on a shared key column in pandas is done with:

Accepted Answer

pd.merge()

Answer

pd.fillna()

Answer

pd.dropna()

Answer

pd.cut()

Question 2

An 'inner join' between two tables returns:

Accepted Answer

Only rows with matching keys in both tables

Answer

All rows from both tables

Answer

All rows from the left table only

Answer

Rows with no matches

Question 3

MICE (Multiple Imputation by Chained Equations) is a method for:

Accepted Answer

Imputing missing values using other features

Answer

Removing duplicate rows

Answer

Scaling features

Answer

Encoding categories

Question 4

Reshaping data from wide to long format in pandas typically uses:

Accepted Answer

pd.melt()

Answer

pd.merge()

Answer

pd.concat()

Answer

pd.dropna()

Question 5

When categorical labels have a natural order (low, medium, high), the best encoding is:

Accepted Answer

Ordinal encoding

Answer

One-hot encoding

Answer

Random encoding

Answer

Mean imputation

Question 6

A 'data dictionary' in a preparation workflow primarily documents:

Accepted Answer

Each field's meaning, type, and allowed values

Answer

The model's accuracy

Answer

Server hardware specs

Answer

Git commit history

Question 7

SMOTE is applied during preparation mainly to address:

Accepted Answer

Class imbalance by synthesizing minority-class samples

Answer

Missing values

Answer

Feature scaling

Answer

Duplicate detection