The Graphite Lab
Sign inBrowse the Catalog

Tools

Anomaly Detector

Surfaces rows in a structured dataset that fall outside expected patterns, each with a reason.

Module
Tools
where it lives

What it does

Surfaces rows in a structured dataset that fall outside expected patterns, each with a reason

Accepts numeric or tabular content and returns the rows that fall outside expected patterns: spikes, drops, missing sequences, and statistical outliers. Each flagged row arrives with a plain-language reason drawn from the data. Works on time series, transaction logs, sensor readings, or any structured dataset with a baseline to compare against.

When you’d reach for it

  • Scanning a batch of technician timesheets for hours logged well outside the typical range for that job type.
  • Reviewing utility consumption readings across a portfolio of units to surface unusual spikes between billing cycles.
  • Running a feed of vendor invoices through outlier detection to catch duplicate amounts or line items that deviate from historical averages.