prompt

Outlier Detection Planner

Recommends an outlier detection approach for a dataset with method choice and handling steps.

VettedUpdated June 2026
The prompt
Plan outlier detection for {{column}} in {{dataset}}, which is {{distribution}} distributed. Recommend a detection method (for example IQR, z-score, or isolation forest) with the reason, the threshold to use, how to confirm a true outlier, and options for handling them. Goal: {{goal}}.
Did it work? Rate this prompt

Variables

columnColumn
datasetDataset
distributionDistribution
goalGoal

Example output

Column: session_duration in events, right-skewed. Method: IQR — robust for skewed data and easy to explain. Threshold: flag values below Q1 - 1.5*IQR or above Q3 + 1.5*IQR. Confirm: check flagged rows for a real cause (a stuck session, a tracking bug) before removing. Handling options: cap at the threshold (winsorize), exclude with a logged reason, or keep and analyze separately. Goal — accurate average session time — favors capping over deletion to avoid bias.

Details

Author

AI Khazna

License

Security

Vetted

Type

prompt

Related assets

More curated picks in Data & Analytics.

Audit before you install

Run any source through our checks - AI visibility, security, performance, and stack detection.

More in Data & Analytics