- Identify the X or predictive, independent variable(s) and the Y or dependent variable to be predicted.
- Determine if the required data of the model is available in the Power BI data model.
In this example, monthly marketing spend from a General Ledger fact table is used to predict monthly internet sales from an internet sales transaction fact table. Simple (single variable) regression models are often insufficient to estimate Y values accurately, but many of the concepts and techniques used in this recipe are applicable to more complex, multiple linear regression models.