Turn raw, messy data into clean, model-ready features using pandas and scikit-learn. You will clean and impute data, encode categoricals, scale numerics, handle outliers, engineer and select features, and assemble it all into a leak-free ColumnTransformer pipeline you can trust on unseen data.
Watch the free preview
Inspecting and cleaning a raw dataset — free to watch, no account needed.
What you'll learn
- Turn raw data into model-ready features with pandas and scikit-learn
- Handle missing data and outliers without distorting the signal
- Encode categorical variables correctly with one-hot and ordinal encoding
- Scale and normalise numeric features and create useful new ones
- Assemble a leak-free ColumnTransformer pipeline that generalises to unseen data
Syllabus
Cleaning Raw Data
Inspecting and cleaning a raw datasetFree preview
Handling missing data
Encoding and Scaling
Encoding categorical variables
Scaling and normalisation
Outliers and Feature Creation
Detecting and handling outliers
Creating useful features
Selection, Leakage and Pipelines
Feature selection and avoiding leakage
Assembling a ColumnTransformer pipeline
Lab — 10 Exercises & Solutions
Exercises 1–5
Exercises 6–10