Loaded Model: Offers fast response times, and includes scheduled reloads. If you can stay under 1GB (post compression) for Power BI Pro users, or the 10GB (post compression) limit for Premium users then the loaded model is the best choice. Direct Query: Directly querying the database is useful if you are analyzing large fact tables that are beyond the Loaded Model limit, or the data refreshes more than once an hour. Good network bandwidth is essential for Direct Query. Live Connect: Similar to Direct Query, but is used when querying data stored in Analysis Services, or HANA. It is highly recommended to leverage Analysis Services to manage row level security in Power BI. Mix Mode: Allows you have some of the tables in your PBI model to be loaded, and others to use direct query. This enables you to load smaller dimensional data in Power BI, and still use direct query for larger fact tables. Common Data Model: Using Dataflows within Power BI Service you can load
data from multiple systems and applications together. You can then create multiple PBI
Datasets from the model which can be leveraged in Power BI reports,
and dashboards.
Power BI Data Sources
The two main questions I hear all the time are “where should I source my data”, and “where should I store my data” for Power BI. The answers depend on the underlying data, and analytical requirements. (Depending on your data sources a variety of PBI storage/connection options are available.)