Use statistical techniques to explore your dataset fields. Common usages:
Fraud detection: by relying on Benford’s Law you can find symptoms of anomalous values in your dataset fields
Normality: find out whether the data in a field is normally distributed. Tests: Anderson-Darling, Jarque-Bera, Z-score.
Outliers: find out whether a field contains any value that differs significantly from the mean. Tests: Grubbs.