Encyclopedia · 176 concepts

Core ML Concepts · beginner · concept 17 of 176

Exploratory Data Analysis (EDA)

The discipline of looking at the data before modeling it: distributions, missing values, outliers, correlations, and the weird rows that reveal how the data was actually collected. An hour of EDA routinely saves a week of debugging a model trained on data you did not understand.

Key terms

DistributionsOutliersMissing valuesCorrelation matrixData profiling

Where you meet it in the real world

The first day of every real ML project; where data leakage and label problems get caught early

Videos

Exploratory Data Analysis

IBM Technology · YouTube