For those not experienced in business intelligence, there is a temptation to treat Power BI as an extension to the existing data source, merely modelling data as it exists in source databases or Excel spreadsheets. At first, this ‘appears’ to work as either fully normalised database tables or Excel-style slabs of data which can be related into Power BI and used to drive visualisations, however for the ‘have a go hero’ of Power BI the modelling stage is often overlooked.
The danger of skipping the modelling stage is although the dashboard may work initially, the data model quickly becomes a problem as the system grows. Wide flat tables will be inefficient to load and hard to cross-reference across datasets and creating horrible many-to-many relationships between data sets for all but the simplest of models.
Alternatively, highly normalised, transaction type tables are inefficient and cumbersome to use within Power BI, resulting in too many small tables and no summarisation.
Star schemas are commonly used to produce optimal models within Power BI. Star schema origins are carried over from enterprise data warehouse where Ralph Kimball has championed their benefits, they proved invaluable for business intelligence modelling.
Star schemas are commonly used to produce optimal models within Power BI. Star schema origins are carried over from enterprise data warehouse where Ralph Kimball has championed their benefits, they proved invaluable for business intelligence modelling.
If you can, starting with a well-modelled environment is always more efficient than converting an already existing implementation. Always try to look at how the business is likely to use the data so a model can be optimised appropriately for the business’s analysis. A bad data model becomes a weighted chain around the implementation, with constant workarounds, poor relationships, and performance.
A good data model will make Power BI fly.
At WB Data Focus we enjoy nothing more than seeing a well-planned Power BI implementation bring data and actionable analytics to life within a company. If you are at the beginning of your Power BI journey or wanting to take make better use of Power BI, feel free to contact us to discuss the next steps.