A decision-focused perspective

An analysis is not complete when the code runs or the model produces output. It is complete when the evidence has been checked, the limitations are understood, the result answers a meaningful question, and the audience can see what the result implies. Throughout the book, technical work is therefore connected to four recurring questions:

1. What decision or understanding is needed?
2. What evidence can support it?
3. How strong is that evidence?
4. What action, if any, is justified?