Outlier Detector

An outlier is an observation unusually distant from the rest of a distribution under a chosen rule. IQR fences and z-scores flag different patterns and rely on different assumptions; a flagged value may be an error, a rare valid event, or evidence that the data need subgroup analysis.

Outlier Detector applies this concept to its defined inputs and workflow. It runs only after a CSV is loaded and you click Detect Outliers.

After the user explicitly loads local data, the tool applies the selected workflow to local CSV file and returns outlier rows, outlier count visual, and outlier CSV. Input structure, labels, missing values, and selected options must match the intended workflow. The output should be compared with the source file and the assumptions of the intended analysis before any transformation is accepted. Outlier detection labels potential outliers; it does not prove that a value is wrong.