admin

he use of separate dataflows based on purpose: staging, transformation, or final. We recommend using composable building blocks to obtain greater reuse, in various combinations, to support specific user requirements. Composable building blocks are described later in this article.

focus of this advanced data preparation scenario is on:

For a dataflow that's intended for reuse, it's common (but not required) that the creator belongs to a centralized team that supports users across organizational boundaries (such as IT, enterprise BI, or the Center of Excellence).

A staging dataflow is created in a workspace that's dedicated to the centralized management of dataflows. A staging dataflow copies the raw data as-is from the source. Few, if any, transformations are applied.

Use of composable building blocks is a design principle that allows you to manage, deploy, and secure system components, and then use them in various combinations. Creating modular, self-contained dataflows that are specific to a purpose is a best practice. They help to achieve data reuse and enterprise scale. Modular dataflows are also easier to manage and test.

Usually, there are many more report creators than dataset creators. These report creators can exist in any area of the organization. Because self-service report creators often need to quickly produce content, a blended approach allows them to focus on producing reports that support timely decision-making without the additional effort of creating a dataset.

The primary objective is for many report creators to reuse centralized shared datasets. To accomplish that, this scenario focuses on decoupling the model development process from the report creation process.

For simplicity, the scenario diagram depicts just one shared dataset. However, it's not usually practical to model all organizational data in a single dataset.

The other extreme is to create a new dataset for every report, as less experienced content creators often do. The goal of managed self-service BI is to find the right balance, leaning toward relatively few datasets and creating new datasets when it makes sense to do so.

hile reports are maintained by subject matter experts in the business units. However, that's not required. For example, this pattern can be adopted by any content creator that wants to achieve reusability.

managed self-service BI is characterized by a blended approach that emphasizes discipline at the core and flexibility at the edge.

However, when the core data architecture doesn't include all data required, dataset creators can extend, personalize, or customize existing shared datasets. New specialized datasets can be created that meet business requirements not met by existing centrally delivered datasets. Importantly, there's no duplication of core data. This usage scenario is called customizable managed self-service BI.

Sometimes self-service creators need to augment an existing dataset with, for instance, additional data that's specific to their department. In this case, they can use DirectQuery connections to Power BI datasets

This feature allows for an ideal balance of self-service enablement while taking advantage of the investment in centrally managed data assets. The scenario diagram depicts a DirectQuery connection