2. Prepare the Data (25–30%)

Every number on a report inherits the quality of the data underneath it. This phase covers the first two segments of the pipeline from Phase 1: getting data in (2.1), checking whether it can be trusted (2.2), and shaping it into the fact and dimension tables a good model needs (2.3). Almost all of it happens in the Power Query Editor, and almost all of it is tested through scenarios: a refresh that fails, a column full of nulls, a flat file that must become a star schema. Read every stem in this phase with one question in mind: what has to be true about this data before a model can use it?

Think of Power Query as a recorded recipe: each applied step replays on every refresh, so a step that hard-codes a column name or a file path breaks the whole refresh when the source changes.

⚠️ Exam Trap: Answers that patch data inside a visual or a DAX measure usually lose to answers that fix it once in Power Query, where every report benefits.