Forum Discussion
Data preparation
JohnUnfiro In general, yes it is a generally accepted best practice to do the data preparation prior to Power BI and to have it pushed as far up the data preparation chain as possible. The only advantage to doing this in Power BI is if you do not have access to the source systems.
Best practice is generally to make sure your data is clean and transformed in the shape you want it before loading it into your Power BI data model.
Regarding summarizing/aggregating, it's a bit more nuanced. In general, you should load data at a granularity that allows you to slice and dice as needed. That is, you want to avoid prematurely aggregating values that you might want to uncombine later. For really big datasets, aggregations may be needed for performance reasons and considerations get more complicated for everything.
Best practice is generally to make sure your data is clean and transformed in the shape you want it before loading it into your Power BI data model.
Regarding summarizing/aggregating, it's a bit more nuanced. In general, you should load data at a granularity that allows you to slice and dice as needed. That is, you want to avoid prematurely aggregating values that you might want to uncombine later. For really big datasets, aggregations may be needed for performance reasons and considerations get more complicated for everything.