Star schema concept in Power BI - Step-by-Step Guide

Introduction

A star schema organizes data into a central fact table connected to several dimension tables. This layout helps you analyze data easily and quickly by simplifying relationships.

When you want to build a sales report that shows total sales by product, region, and time.

When you need to analyze customer behavior using different categories like age, location, and purchase history.

When your data has clear facts (numbers) and descriptive details (categories) that relate to those facts.

When you want to improve report performance by reducing complex joins.

When you want to create a simple and understandable data model for business users.

Steps

Step 1: Import your main data table

- Power BI Desktop > Home tab > Get Data

The fact table with numeric data appears in the Fields pane

💡 The fact table usually contains sales amounts, quantities, or counts

Step 2: Import dimension tables

- Power BI Desktop > Home tab > Get Data

Dimension tables with descriptive data appear in the Fields pane

💡 Dimension tables include product names, dates, or customer info

Step 3: Create relationships between fact and dimension tables

- Model view > Drag the key column from dimension table to matching column in fact table

Lines appear connecting the fact table to each dimension table

💡 Use one-to-many relationships with the dimension table on the 'one' side

Step 4: Verify relationship directions and cardinality

- Model view > Select relationship line > Properties pane

Relationships show single direction and correct cardinality

💡 Single direction filtering improves performance and avoids ambiguity

Step 5: Use fields from dimension tables to slice and filter your reports

- Report view > Fields pane > Drag dimension fields to Filters or Slicers

Report visuals update based on selected dimension filters

💡 This shows how star schema supports easy data exploration

Before vs After

Before

Fact and dimension tables are imported but not connected; report shows all data without filtering

After

Fact and dimension tables are connected in a star schema; report filters and slices work correctly and efficiently

Settings Reference

Cardinality

📍 Model view > Relationship properties

Defines how tables relate to each other, usually dimension to fact is one to many

Default: One to many

Cross filter direction

📍 Model view > Relationship properties

Controls how filters flow between tables; single is best for star schema

Default: Single

Active relationship

📍 Model view > Relationship properties

Determines which relationship Power BI uses by default in calculations

Default: Active

Common Mistakes

Creating many-to-many relationships between fact and dimension tables

This causes ambiguous filters and slows down reports

Ensure dimension tables have unique keys and use one-to-many relationships

Using bidirectional cross filtering in star schema

It can cause unexpected filter behavior and performance issues

Use single direction filtering from dimension to fact table

Summary

Star schema organizes data with one fact table connected to multiple dimension tables.

It simplifies data relationships and improves report performance.

Use one-to-many relationships with single direction filtering for best results.

Practice

1. What is the main characteristic of a star schema in Power BI?

easy

A. A single table with all data combined

B. Multiple fact tables connected in a chain

C. One central fact table connected to multiple dimension tables

D. Only dimension tables without any fact table

Solution

1. Step 1: Understand star schema structure

   A star schema has one main fact table that stores measurable data.

2. Step 2: Identify dimension tables role

   Dimension tables provide descriptive attributes and connect to the fact table.

3. Final Answer:

   One central fact table connected to multiple dimension tables -> Option C

4. Quick Check:

   Star schema = One fact + many dimensions [OK]

Hint: Remember star schema looks like a star with one center [OK]

Common Mistakes:

• Thinking multiple fact tables form a star schema
• Confusing dimension tables as fact tables
• Assuming star schema has no fact table

2. Which of the following is the correct way to describe the relationship in a star schema?

easy

A. One-to-many from dimension tables to fact table

B. Many-to-many between fact and dimension tables

C. One-to-many from fact table to dimension tables

D. One-to-one between fact and dimension tables

Solution

1. Step 1: Identify relationship direction

   Dimension tables filter many rows in the fact table, so one dimension row relates to many fact rows.

2. Step 2: Confirm relationship type

   This means a one-to-many relationship from dimension to fact table.

3. Final Answer:

   One-to-many from dimension tables to fact table -> Option A

4. Quick Check:

   Dimension to fact = one-to-many [OK]

Hint: Dimension filters many facts, so one-to-many from dimension [OK]

Common Mistakes:

• Reversing the relationship direction
• Choosing many-to-many which is not typical in star schema
• Assuming one-to-one relationships

3. Given a star schema with a Sales fact table and a Product dimension table, what happens if you slice sales by product category?

medium

A. The report shows total sales without any filtering

B. Sales values are grouped by product category from the dimension table

C. Sales values are ignored because categories are not in the fact table

D. An error occurs because dimension tables cannot filter facts

Solution

1. Step 1: Understand slicing in star schema

   Slicing by product category uses the Product dimension to filter the Sales fact table.

2. Step 2: Effect on sales values

   Sales are grouped and summed by the selected product categories.

3. Final Answer:

   Sales values are grouped by product category from the dimension table -> Option B

4. Quick Check:

   Slicing by dimension groups facts [OK]

Hint: Dimensions slice facts to group data in reports [OK]

Common Mistakes:

• Thinking dimension filters do not affect fact data
• Assuming error occurs when slicing by dimension
• Believing sales values are ignored

4. You created a star schema but your report shows incorrect totals when filtering by a dimension. What is the most likely cause?

medium

A. The report uses a pie chart instead of a bar chart

B. The fact table has duplicate rows

C. The dimension table has too many columns

D. The relationship between fact and dimension tables is inactive or missing

Solution

1. Step 1: Check relationships in star schema

   Inactive or missing relationships prevent filtering from dimension to fact table.

2. Step 2: Impact on report totals

   Without active relationships, filters do not apply correctly, causing wrong totals.

3. Final Answer:

   The relationship between fact and dimension tables is inactive or missing -> Option D

4. Quick Check:

   Active relationships needed for correct filtering [OK]

Hint: Verify relationships are active to fix filtering issues [OK]

Common Mistakes:

• Blaming duplicate rows without checking relationships
• Thinking chart type affects data filtering
• Ignoring relationship status

5. You want to optimize a Power BI model with sales data. Which star schema design improves query speed and report clarity?

hard

A. One large fact table with keys to small, focused dimension tables

B. Multiple fact tables with duplicated dimension data inside each

C. One fact table with all dimension attributes merged inside it

D. Only dimension tables with no fact table

Solution

1. Step 1: Understand star schema optimization

   A large fact table with keys to small dimension tables reduces data duplication and improves performance.

2. Step 2: Avoid poor designs

   Duplicating dimension data or merging all attributes in fact table causes slow queries and confusion.

3. Final Answer:

   One large fact table with keys to small, focused dimension tables -> Option A

4. Quick Check:

   Star schema = big fact + small dimensions [OK]

Hint: Keep fact big and dimensions small for best speed [OK]

Common Mistakes:

• Duplicating dimension data in fact tables
• Merging all data into one table
• Removing the fact table entirely