Check data quality
Review the source, collection method, missing values, sample size, and potential selection bias.
INSIGHTS / DATA ANALYSIS
Historical data needs context.
A record of past events can describe what occurred, but it may not explain why it occurred or what will happen next.
THE ESSENTIALS
Built on clear principles.
Separate description from prediction
Descriptive summaries characterize observations; predictions require additional assumptions and validation.
Report uncertainty
Explain the limits of the data and avoid presenting estimates as guarantees.
KEEP EXPLORING
Curiosity is a good place to start.
Browse the learning center for clear explanations of algorithms, probability, data, and digital safety.
Visit the learning center