Data with mistakes, missing parts, or facts that do not match.
Dirty data is a takeout order with "no pickles" marked "extra pickles." The cook can be great. Your burger is still doomed.
AI can learn from dirty data. Then hiring tools or fraud checks may make bad calls.
AI-Ready Data
Dirty data needs cleaning before it becomes AI-Ready Data.
Data Labeling
Wrong or uneven labels are a common source of dirty data.
Data Governance
Data Governance uses rules to find and fix dirty data.