C6PythonIntermediateAbout 120 min to complete
Missing Values
Defensible handling of missing data — and the ability to tell when imputation is honest and when it is fabrication.
Notebook
About 120 minutes to complete
Enrol to read the notebook and download it to run.
What this module covers
Defensible handling of missing data — and the ability to tell when imputation is honest and when it is fabrication.
By the end of it
- Detect and visualise missingness patterns across a dataset
- Distinguish MCAR, MAR, and MNAR, and explain why the distinction matters
- Test whether missingness correlates with other variables
- Choose between deletion, simple imputation, and model-based imputation
- Apply forward/backward fill and interpolation appropriately to time series
- Use sklearn's SimpleImputer and KNNImputer correctly
- Quantify how imputation distorts variance and correlations
- Report missingness honestly in any analysis