To have a better understanding of Power BI, we can divide the architecture into three parts or phases:

1. Data Integration

In Power BI, we can import data from different kinds of data sources in different formats. In the data integration step, Power BI brings data together (extracted) from different data sources and converts it into a standard format. After data is integrated into Power BI, it is stored in a common storage area known as the staging area.

2. Data Processing

Once Power BI integrates and stores data at a secure place, the raw data requires some processing. Several processing or cleansing operations transform the raw data such as removing redundant values, etc. Later, we apply relevant business rules on the processed data that transforms it according to our business needs. This transformed data is loaded into the data warehouses. This completes a full process of ETL.

3. Data Presentation

In this final phase, the processed data moves from the warehouse and goes into the Power BI platforms like Power BI Desktop to create reports, dashboards, and scorecards. Power BI offers a wide range of visualizations. We can also import custom visualization from the marketplace. From the report development platforms, we can publish the reports on the web or mobile apps to share it with other business users.