SECAI Core · phase 3 of 15
Data cleaning
The process of ensuring that the data used in analysis is of a high-enough quality that it gives decision makers confidence in findings and the practice of transforming non-conforming data into acceptable shape.
The Explain card
- Plain English
- Data cleaning makes sure data is good enough to trust: fixing errors, filling or removing bad values, and reshaping non-conforming records into an acceptable form.
- Example
- Before training a fraud model, analysts strip out test transactions, correct mistyped currency codes and drop records with impossible dates.
- Why it matters
- Dirty data produces unreliable models, but cleaning is also a security step. It is where poisoned outliers, injected instructions and leaked secrets can be caught before they are learned.
- Hook
- Garbage in, garbage learned; clean before you train.
Where it sits in the deck
Phase 3: Data Fundamentals: Types, Pipelines, and Preparation
Models are only as good as their data — understand the raw material and the engineering that shapes it before covering how it flows or breaks.
