EDA100% Active LaboratoryLAB
Categorical Encoding (One-Hot, Ordinal, Target) Lab
Interactive categorical variable transformer demonstrating dummy variable traps, high cardinality dimensionality explosions, and target leakage.
EDA • VISUALIZATIONThe Mean Lies & Tukey's IQR Outlier Laboratory
EDA • DESCRIPTIVE STATISTICSDistribution & Outlier Sensitivity
"The Mean Lies" & Tukey's IQR Outlier Laboratory
Outliers Found:1 points
IQR Fence Multiplier:1.5 × IQR
Mean (μ)
36.67
Distorted by extreme values
Median (Q2)
28.00
Robust to outliers!
IQR (Q3 - Q1)
4.00
Middle 50% spread
Tukey Fences
[20.0, 36.0]
Values outside are outliers
1D Dot Plot with Fences:
Crucial EDA Principle:
The Mean is sensitive to extreme values because it minimizes squared errors ($\sum (x_i - \mu)^2$). A single massive outlier drags the mean significantly toward itself. In contrast, the Median and IQR are rank-based order statistics that remain completely unaffected by the extreme magnitude of outliers.
Want a step-by-step breakdown from our 4 AI Engineering Agents?
Mastery Standard: Complete all 3 tasks in the Lab Protocol to earn verified mastery credits.
Study Full Unit Notes →