outlier

Outlier

A data point that falls unusually far from the rest of the dataset — either an extreme legitimate value or a data error.

When you'd see it: Any time the mean and median diverge significantly — outliers are probably pulling the mean. Also in fraud detection (outlier transactions), quality control (outlier measurements), and data cleaning (outlier values that indicate input errors).

Why it matters: Outliers can be the most important data points in a dataset (fraud, equipment failure, exceptional customers) or they can be errors (typos, unit mismatches). Deciding which case applies before removing or keeping them is one of the most consequential judgment calls in data analysis.

Common mistakes: Automatically removing outliers to 'clean' the data. Outlier removal can destroy your ability to detect the exact phenomena you care about. Always investigate whether an outlier is a measurement error or a real observation before touching it.

Related mean, median, standard deviation, normal distribution