Power BI Dataflows and Data Lakes

[Anonymous]

Hello, 

Just reading the November update for Power BI which indicates the following:

• Dataflows can be configured to store the data in the customer’s Azure Data Lake Storage Gen2 instance, fueling collaboration across roles. Business analysts can seamlessly operate on data stored in Azure Data Lake Storage, taking advantage of its scale, performance, and security. Meanwhile, data engineers and data scientists can extend insights with advanced analytics and AI from complementary Azure Data Services like Azure Machine Learning, Azure Databricks, and Azure SQL Data Warehouse.
• Dataflows support the Microsoft Common Data Model, giving organizations the ability to leverage a standardized and extensible collection of data schemas (entities, attributes and relationships)

Interested in the forum's perspective - is there still a place for Analysis Services (e.g. SSAS Tabular/MDX) as a primary source of information for Power BI reports? Or will dependence on Analysis Services diminish in favour of the above technologies - e.g. PBI data modelling, PBI data service, MS common data model, SQL DWH etc?

Thanks, 

pbix

[v-yuta-msft · Accepted solution]

Hi pbix,

It seems to be a deep topic about relationship and difference between online data process and traditional ETL. I think both are still useful in specific senarios. For power bi datasource, I would recommend you to try data flow because dataflow is a suite self-service low-code/no-code features and capabilities for business analysts to easily process and unify their data and store it in Azure-based data-lake storage.  With these new capabilities, Power BI offers a solution for any business need—whether you want to prep your data with ease, using a familiar built-in Power Query experience, or to leverage the full Azure stack for more advanced use-cases.

Also, in other senario, the comparsion is also meaningful, I would recommend you to refer to blog below:

https://www.jamesserra.com/archive/2017/12/is-the-traditional-data-warehouse-dead/.

Regards,

Jimmy Tao