# Foundations knowledge check

Ten questions on finding, reading and trusting data in Collate. You need 8 correct to pass. You can retake this as many times as you like, and you will see an explanation for every answer.

Question 1

1

A table in Collate has a clear, well written description. What does that tell you on its own?

The description was written and verified by the data ownerSomething described it, either a person or an agent, and you cannot tell which without checkingThe table has passed its quality testsThe table is approved for use in reporting

Question 2

2

You are about to deprecate a table. Which direction of lineage answers your question?

Upstream, to see where its data comes fromNeither, lineage only covers pipelinesDownstream, to see what depends on itBoth directions give the same answer

Question 3

3

A search for "customer" returns 28 tables plus dashboards and data models. What is the most reliable way to narrow it to the sensitive customer tables?

Filter the results by tag, such as a sensitivity classificationSort the results by popularity and take the top fewSearch again using a longer, more specific keywordRestrict the search to a single source system

Question 4

4

An asset has no data quality tests attached. How should you read that?

The data is known to be good, so no tests were neededThe data is unmeasured, and you have no evidence either wayTests are only applied to sensitive data, so this is expectedThe data is known to be bad

Question 5

5

Which failure mode does a freshness test catch that a schema or null check would miss?

Duplicate primary keysA column being renamed upstreamA value falling outside its expected rangeA pipeline that stopped running, leaving data that looks normal but is out of date

Question 6

6

You open the glossary term "Customer Lifetime Value". What does that give you beyond the definition itself?

The row count of every asset that implements itA guarantee that every asset implementing it is correctThe assets that implement it, along with its owners, domain and reviewersPermission to edit those assets

Question 7

7

You spot a wrong description on a table you do not own and have no rights to edit. What is the correct action?

Leave a comment and hope the owner notices itRequest the change, proposing new wording and naming an assignee to review and merge itRaise a ticket with the platform teamNothing, since only the owner can judge whether a description is wrong

Question 8

8

Two dashboards report different revenue figures. Which is most likely to resolve the discrepancy?

Column level lineageTable level lineageThe asset descriptionThe row count of each source table

Question 9

9

You need a dataset for a regulatory filing. It has passing tests, a named owner, and complete lineage. Is it fit for purpose?

No, regulatory work can never use data from the platformYes, those three signals are sufficient for any useYes, provided the owner is in the data teamNo, fit for purpose depends on what you need it for, so you still have to judge against your own requirements

Question 10

10

Where does the metadata in Collate come from?

Connectors read it from source systems on a schedule, and agents enrich what they findStewards enter it manually through formsIt is uploaded in bulk when the platform is first installedIt is derived entirely from query logs

Check my answers0 of 10 answered