GritBot is an automatic tool that tries to find anomalies in data as a precursor to data mining. It can be thought of as an autonomous data quality auditor that hunts for records having "surprising" values of nominal (discrete) and/or numeric (continuous) attributes.