Garbage Data, And What To Do About Them
summary
Framework defines garbage data as data that are not useful for a stated application and gives concrete failure modes: illegible handwriting, storage corruption, mass default-value corruption, and deliberate tampering. Costs are quantified as recurring cloud and retention bills plus hidden opportunity, people and decision costs that can cascade into catastrophic losses. Recommended actions are to destroy or quarantine garbage data, flag it prominently, be honest and empathetic with stakeholders, and treat projects as data projects so teams can validate datasets before analysis.
AI-written summary · about
chapters (4)
- 0:00The problem of garbage data
- 3:37Identifying garbage data
- 8:47The costs of bad data
- 12:20Taking action and best practices
related talks
16mGarbage Data, And What To Do About Them
43mThe Death of Data: Risk, Retention, and Rot
41mData Quality Management: The Good, The Bad, and The Messy
44mPreventing, Diagnosing & Curing Bad Data
48mControlling a Data Footprint – How to Build a Data Disposition Framework
24mData’s Rotten Scoundrels – How To Avoid Becoming One
35mDesigning Data Governance Framework
01h 00mBig Data, Bad Analogies