Data quality checks
Too good to be true
Applications jump 38% overnight. A test on the total decides whether anyone acts on the number.
Topics › Data quality and change › When things go wrong
Silent change: why a change needs both sides
A new enrolment status reaches the data overnight, and each side saw only half of the change. See how a test and a data contract make the next change visible to both sides.
Builds on The Inner Life of Data · Refining with dbt and Gold
Watch the film
Press Play, or pick a chapter. Turn on Pause and think to stop for one question at the end of chapters.
Download the videoCaption filesRead the story outline
How to read the film
Light is data, as in The Inner Life of Data. Each person's outline shows their side: cyan for technical, gold for business. Amber marks a warning, or something new that waits for a decision; red marks a test that stopped the build. The university, people and numbers are fictional. The narration is a synthetic voice.
English captions are on by default; turn them off with Captions. The video you download has none on the picture: play it with its caption file, in English or Spanish.
Where next?
You've seen one change from both sides. Now take it further.
Think it through
The four Pause and think questions, for a class or a team. Open one to see the answer.
Go further
Data modelling
A Sharper Sketch
Two trusted numbers disagree. When does a simple model need more precision, and how do you use a reference model without copying it?
Start here
The Inner Life of Data
Follow one enrolment from a tap on a phone to a decision, through Databricks and dbt.