Overview
Missing data is not a nuisance to be filled and forgotten. It is a signal. The pattern of missingness — whether random, systematic, or structurally determined — dictates which imputation method is valid and which will silently corrupt your analysis. A mean imputation on data that is missing not at random doesn't clean your dataset. It introduces bias you can't see.
Most data cleaning guides treat imputation as a single decision: pick a method, apply it, move on. The reality is that different columns in the same dataset may require different imputation strategies based on their missingness mechanism, their distribution, their relationship to other variables, and their role in the downstream analysis.
The Missing Data Imputation Strategy Prompt generates a complete imputation framework: missingness mechanism diagnosis, method selection logic by variable type and mechanism, implementation specifications, and a validation protocol that detects whether imputation introduced bias.
What you get: - Missingness mechanism diagnosis (MCAR, MAR, MNAR) per variable - Method selection matrix by mechanism and variable type - Implementation specifications for each chosen method - Pre/post imputation validation protocol - Downstream impact assessment for model and analysis integrity
Built for: data scientists, analysts, and ML engineers who need defensible imputation decisions, not default fills.