Scenario: Two teams argue because customer_id means different things. What prevents this best?
contractsScenario: Missing values are treated as zero. What failure can this create?
qualityScenario: A field is an identifier but is used numerically in a model. Why is that risky?
modellingScenario: A metric improves after a pipeline change, but trust drops. What should you check first?
metricsScenario: A dashboard looks great but decisions are wrong. What data risk is most likely?
aggregationScenario: Null and empty string are treated as the same. Why is this risky?
qualityScenario: A dataset contains names and emails. What should happen first?
privacyA data contract should include:
contractsScenario: A KPI changed unexpectedly. What is the fastest way to narrow the cause?
lineageScenario: A dataset is used for decisions but has no owner. What is the risk?
qualityScenario: A metric is used for bonuses. What should you do first to keep it credible?
metricsScenario: Analysts want access to a sensitive dataset. What is the most defensible default?
privacy