ColumnTransformer for mixed types in ML Python - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the purpose of ColumnTransformer in machine learning?

ColumnTransformer helps apply different data transformations to different columns in a dataset. This is useful when columns have mixed types, like numbers and categories.

beginner

How does ColumnTransformer handle numeric and categorical columns differently?

It allows you to specify separate transformers for numeric columns (like scaling) and categorical columns (like one-hot encoding) in one step.

intermediate

Why is it better to use ColumnTransformer instead of transforming columns separately?

Using ColumnTransformer keeps the preprocessing organized, avoids mistakes, and integrates well with machine learning pipelines.

intermediate

What happens if you don't specify remainder='passthrough' in ColumnTransformer?

Columns not listed in transformers are dropped by default. Using remainder='passthrough' keeps those columns unchanged.

beginner

Give an example of transformers used for numeric and categorical data in ColumnTransformer.

Numeric: StandardScaler() to scale numbers.<br>Categorical: OneHotEncoder() to convert categories into binary columns.

What does ColumnTransformer do?

A Splits data into training and testing sets

B Trains multiple models at once

C Visualizes data distributions

D Applies different transformations to different columns

ColumnTransformer applies different preprocessing steps to specified columns.

Which transformer is commonly used for numeric columns?

A StandardScaler

B OneHotEncoder

C LabelEncoder

D CountVectorizer

StandardScaler scales numeric data to have mean 0 and variance 1.

What does OneHotEncoder do for categorical data?

A Converts categories into numbers 1, 2, 3...

B Removes missing values

C Creates binary columns for each category

D Scales categories between 0 and 1

OneHotEncoder creates new columns with 0 or 1 to represent categories.

If you want to keep columns not transformed by ColumnTransformer, what parameter do you use?

A remainder='passthrough'

B keep='all'

C drop=False

D preserve=True

remainder='passthrough' keeps columns not listed in transformers unchanged.

Why is using ColumnTransformer helpful in a machine learning pipeline?

A It automatically selects the best model

B It organizes preprocessing for mixed data types

C It increases model accuracy by itself

D It visualizes the data

ColumnTransformer helps organize different preprocessing steps for different columns.

Explain how ColumnTransformer helps when your dataset has both numeric and categorical columns.

Think about how you prepare numbers and categories differently.

Describe what happens if you do not set the remainder parameter in ColumnTransformer.

Consider what happens to columns you forget to mention.