Prepping data for tabular models
With multidimensional models, a star schema is required in the underlying data source. However, with tabular models, a star schema is not required. This means that data preparation is not as clear as it is with multidimensional models. In this section, we will explore some key considerations that are involved when preparing data for tabular models.
Contrasting self-service and managed deployments
Tabular model designs have their origins in self-service technologies such as Power BI and Excel. Why does this matter? Because well-designed dimensional models still perform better and are easier to develop solutions for. Self-service models often focus only on the immediate business need and not on lasting performance or growth. When the number of consumers of an analytics model is one or just a few, the impact is minimal. However, when scaling the models beyond a limited set of users, performance and usability become key considerations in design.
...
