SECAI Core · phase 4 of 15 · noun

Data lineage

A record of the origin, movement, and transformation of data as it flows through systems and processes.

The Explain card

Plain English: Data lineage records how data moves and changes as it flows through systems and processes, from its origin to its final use.
Example: A lineage graph shows that a customer field in a training table was copied from the CRM, joined with support tickets, anonymised, then exported to the model training bucket.
Why it matters: Lineage lets you trace a bad prediction back to its source, find every model touched by a poisoned dataset, and prove to auditors which personal data ended up where.
Hook: Follow the river upstream to find who muddied it.

How the word is built, where it came from, and what it sits beside in memory.

In a sentence

In the feature store, data lineage for the fraud-score column traces the CRM extract through the Spark job that wrote last night's training table.

Why these words

data Latin data, things given, plural of datum supplies the given-facts half, so descent is of records rather than of people
lineage Old French lignage, line of descent, from ligne, line, from Latin linea, linen thread gives the phrase its descent word, so each join can name a parent row

Where it came from

Origin: English computing collocation from Latin data (things given, plural of datum) and Old French lignage (line of descent, from ligne, line, from Latin linea, linen thread)
Entered the language: 1990s
What changed: Data had meant things given, then facts granted in an argument, then the records a machine keeps. Lineage entered English around 1300 as a bloodline, then a pedigree more generally. Warehouse papers of the 1990s borrowed that descent word for records that could name a parent after each join, and AI later asked the same pedigree of a training row.

How it is spelled

Pattern: data stays the Latin plural of datum, even when English treats it as a mass noun
Pattern: -age nouns from French name a collective or result
Pattern: the medial e marks an extra syllable, lin-e-age, not the two-syllable linage
Breaks the pattern: the noun phrase stays two words; the hyphen appears only when it modifies another noun
Breaks the pattern: data looks plural but computing English often treats it as a mass noun
Breaks the pattern: lineage looks like line plus age, but the first vowel is usually short, not the diphthong of line

Spelled like

Broken into chunks

What it sits beside

Same subject

Same shape

Data compounds

Lineage compounds

Lineage and custody

Phase 4: Data Integrity, Governance, and Provenance

Once you know what data is, learn the principles that keep it trustworthy, traceable, and minimized throughout its lifecycle.