Spotting outliers

An outlier is a value far outside the usual range: a customer aged 212, an order of 10,000 units, a negative delivery time. Some are errors, others are real and important events.

A single outlier can shift an average, trigger a false alert or skew a model. Deleting every unusual value is just as risky: it can hide fraud, a sensor failure or the start of a trend.

Good practice is to flag rather than delete: compare the value with business rules and with the rest of the data, look for the cause, then decide to correct, keep or exclude it, and record why.

In the reports of your games, every answer linked to this skill counts: the skill is shown as acquired from 80% of correct answers, and in progress from 50%.