Power BI connects, prepares, models and visualizes data. This introduction focuses on a step that strongly affects whether a report is useful: building a clear data model before designing dashboards. The session uses retail examples such as sales, stock levels, customer behavior, product trends, recommendations and pricing.
Start with the business question
Before importing data, decide what the report should help someone understand. A sales report may need to answer how revenue changes by product, date, store or customer group. Clear questions help determine which tables, measures and filters belong in the model.
Fact and dimension tables
A fact table records events or measurements, such as order lines, sales amounts or stock balances. Dimension tables describe the context used to filter and group those facts—for example, product, customer, location and date. Keep the grain of the fact table consistent: a row might represent one product on one order line, rather than a mixture of daily and monthly totals.
Why a star schema helps
In a star schema, dimension tables connect to a central fact table through relationships, commonly one-to-many. Dimensions support filtering and grouping; facts support summarization. A well-designed model can make measures easier to write and reports easier for others to use. Microsoft’s guidance recommends separating fact and dimension roles and keeping the fact-table grain consistent.
Relationships, filters and many-to-many cases
Relationships define how filters flow between tables. One-to-many relationships are common. Many-to-many cases need extra care; a bridge table can model a relationship between dimensions where each item may relate to several others. Do not add a relationship only to make a visual display data—check that its direction, cardinality and keys represent the business logic.
Turn the model into retail analysis
- Use a date dimension to compare sales over time.
- Use product and location dimensions to find which categories or stores contribute to results.
- Compare sales with inventory while keeping each table’s grain clear.
- Review customer and recommendation metrics with a defined period and denominator.
- Document pricing measures so a change in revenue is not confused with a change in volume.
Continue learning
Power BI features and interface labels can change. For current instructions, use Microsoft Learn’s documentation and training. This article summarizes the session’s introductory topics and is not a complete step-by-step manual.