Analysis of statistical distribution, detection of systemic bias, verification of data leakage between train/test splits, and anomaly control in machine learning datasets. Covers completeness, class balance, label quality, and identification of outliers or duplicates that could compromise the generalization and fairness of the trained model.