Data lineage completeness checker
Anyone who buys, audits or trains on your data will ask the same question: can you prove where it came from and what was done to it? List your main datasets and tick what you can document for each, from the source system and legal basis to change history and retention. The checker scores every dataset, flags the gaps a buyer would treat as deal-breakers and gives you a fix list in order of importance.
Score by dataset
Each attribute carries a weight; the legal basis and source together are a third of the score.
| Dataset | Score | Missing | Status |
|---|
Fix list
Critical gaps first, then the gaps that affect the most datasets.
How the lineage score works
The eight attributes are the questions buyers, auditors and privacy regulators ask about a dataset. Each one has a weight based on how much it matters when data changes hands.
Weight each attribute
Legal basis 20, source 15, changes 15, and owner, dates, versions, access log and retention 10 each, for a total of 100.
Score each dataset
A dataset scores the sum of the weights you ticked.
dataset score = Σ weight of each ticked attribute
Average across datasets
The overall score is the plain average of every named dataset.
Give a status
A dataset missing its legal basis or source is marked "Blocks a sale" whatever its score. Otherwise 85 or more is audit-ready, 60 to 84 needs work and below 60 is weak.
Questions about data lineage
What is data lineage?
Data lineage is the record of where a dataset came from, what happened to it along the way and who is responsible for it. Complete lineage lets anyone trace a record back to the system and moment it was created, and see every change made to it since.
Why do data buyers care about lineage?
A buyer who trains an AI model on your data takes on legal risk if it later turns out you had no right to share it. Lineage is how you prove that right. Buyers ask for it during due diligence, and gaps either lower the price or stop the deal.
Which lineage gaps are the most serious?
A missing legal basis and an unknown source are the two that block deals, which is why the checker marks them as critical. Without them nobody can confirm the data was collected lawfully. Missing access logs or version history are weaknesses, but they can usually be fixed going forward.
How do I fix lineage for data I already have?
Start with a simple register: one row per dataset with its source system, owner, date range, legal basis and retention rule. Then keep the original export untouched, do all cleaning on copies and write each step into a change log. That covers most of what a buyer or auditor will ask.
Do I need special software for data lineage?
Not for a sale or an audit of a handful of datasets. A spreadsheet register, a change log and a folder of contracts and consent records are enough for most small and mid-size companies. Lineage software is useful when data moves through many automated pipelines.