Why should you always do group-wise EDA?

medium

Answer

$Whole-population statistics can hide subgroup patterns that matter for the model: (1) Simpson's paradox — trends flip when you disaggregate.$
$(2) Fairness — model performance may differ by demographic group.$
$(3) Missingness patterns may vary by group.$
$(4) Different scale / variance per group.$
$(5) Distinct outlier populations.$
$Practice: pick 2-3 key categorical variables (region, cohort, product), replot univariate + bivariate patterns by group.$
$Reveals lots of hidden structure.$

Check yourself — multiple choice

Group by never useful
Subgroup analysis reveals Simpson's paradox / fairness gaps / group-specific missing patterns / distinct outliers
Always identical
Same as global

Group-wise EDA: catches Simpson's paradox, fairness gaps, per-group patterns.

#eda

215 interview questions in this topic.