Reducing model size in Power BI - Real Business Scenario

Scenario Mode

👤 Your Role: You are a Power BI developer at a retail company.

📋 Request: Your manager wants you to reduce the size of the current sales data model to improve report performance and loading speed.

📊 Data: You have a sales dataset with columns: OrderID, ProductID, ProductName, Category, SalesAmount, OrderDate, CustomerID, CustomerName, Region, and some calculated columns.

🎯 Deliverable: A Power BI report with a reduced data model size by removing unnecessary columns, optimizing data types, and using aggregations where possible.

Sample Data

OrderID	ProductID	ProductName	Category	SalesAmount	OrderDate	CustomerID	CustomerName	Region
1001	200	Chair	Furniture	150	2024-01-05	3001	John Doe	East
1002	201	Table	Furniture	300	2024-01-06	3002	Jane Smith	West
1003	202	Lamp	Lighting	80	2024-01-07	3003	Bob Brown	North
1004	203	Sofa	Furniture	500	2024-01-08	3004	Alice Green	South
1005	204	Desk	Furniture	250	2024-01-09	3005	Tom White	East
1006	205	Ceiling Light	Lighting	120	2024-01-10	3006	Emma Black	West
1007	206	Bookshelf	Furniture	200	2024-01-11	3007	Chris Blue	North
1008	207	Floor Lamp	Lighting	90	2024-01-12	3008	Pat Yellow	South

1

Step 1: Remove unnecessary columns from the data model that are not used in reports or calculations.

Remove columns: CustomerName, ProductName

Expected Result

Data model size decreases by removing text-heavy columns.

2

Step 2: Change data types to more efficient types where possible.

Change OrderDate column type to Date instead of DateTime; change SalesAmount to Decimal with 2 decimal places.

Expected Result

Data model size reduces by optimizing data types.

3

Step 3: Create a summary table to aggregate sales by Category and Region to reduce row count.

Use DAX: SalesSummary = SUMMARIZE('Sales', 'Sales'[Category], 'Sales'[Region], "TotalSales", SUM('Sales'[SalesAmount]))

Expected Result

Aggregated table with fewer rows and smaller size.

4

Step 4: Disable loading of the original detailed sales table to keep only the summary table in the model.

In Power Query, right-click original Sales table and select 'Disable Load'.

Expected Result

Model size reduces significantly by keeping only aggregated data.

5

Step 5: Create a measure to calculate total sales from the summary table for use in reports.

Total Sales = SUM(SalesSummary[TotalSales])

Expected Result

Reports show correct total sales using smaller model.

Final Result

-----------------------------------
| Category | Region | Total Sales |
|----------|--------|-------------|
|Furniture | East   | 400         |
|Furniture | West   | 300         |
|Furniture | North  | 200         |
|Furniture | South  | 500         |
|Lighting  | East   | 0           |
|Lighting  | West   | 120         |
|Lighting  | North  | 80          |
|Lighting  | South  | 90          |
-----------------------------------

✓Removing unused text columns reduced model size.

✓Changing data types to more efficient ones saved space.

✓Aggregating data into a summary table greatly reduced row count.

✓Disabling load of detailed data improved report performance.

✓Final report shows total sales by category and region correctly.

Bonus Challenge

Implement incremental data refresh to load only new sales data each day to keep the model size small over time.

Show Hint

Use Power BI incremental refresh settings with a date filter on the OrderDate column.

Practice

1. Which of the following is a simple way to reduce the size of a Power BI data model?

easy

A. Import duplicate tables for safety

B. Add more calculated columns

C. Remove columns that are not used in reports

D. Enable auto date/time for all tables

Solution

1. Step 1: Identify unused data

   Unused columns take up space without adding value to reports.

2. Step 2: Remove unused columns

   Deleting these columns reduces the data model size and improves performance.

3. Final Answer:

   Remove columns that are not used in reports -> Option C

4. Quick Check:

   Removing unused data reduces model size [OK]

Hint: Delete unused columns to shrink model size quickly [OK]

Common Mistakes:

• Keeping all columns regardless of usage
• Adding calculated columns unnecessarily
• Enabling auto date/time increases model size

2. Which syntax correctly disables the auto date/time feature in Power BI using the Options menu?

easy

A. Go to File > Options > Data Load > uncheck 'Auto Date/Time'

B. Use DAX: DISABLE AUTODATE()

C. Right-click table > Disable Auto Date

D. In Power Query, remove date columns

Solution

1. Step 1: Locate auto date/time setting

   It is found under File > Options > Data Load in Power BI Desktop.

2. Step 2: Disable the feature

   Uncheck the 'Auto Date/Time' option to stop Power BI from creating hidden date tables.

3. Final Answer:

   Go to File > Options > Data Load > uncheck 'Auto Date/Time' -> Option A

4. Quick Check:

   Disabling auto date/time reduces hidden data [OK]

Hint: Disable auto date/time in Options to reduce hidden tables [OK]

Common Mistakes:

• Trying to disable auto date/time with DAX
• Removing date columns instead of disabling feature
• Looking for option in Power Query instead of Options menu

3. Given a table with a column 'SalesAmount' stored as Decimal Number, which change reduces model size the most without losing data precision?

medium

A. Keep 'SalesAmount' as Decimal Number

B. Change 'SalesAmount' data type to Whole Number

C. Change 'SalesAmount' data type to Text

D. Change 'SalesAmount' data type to Fixed Decimal Number

Solution

1. Step 1: Understand data types and size

   Decimal Number uses more space; Fixed Decimal Number stores decimals efficiently.

2. Step 2: Choose Fixed Decimal Number

   This type reduces size while keeping decimal precision, unlike Whole Number or Text.

3. Final Answer:

   Change 'SalesAmount' data type to Fixed Decimal Number -> Option D

4. Quick Check:

   Fixed Decimal Number reduces size with decimals [OK]

Hint: Use Fixed Decimal Number for decimals to save space [OK]

Common Mistakes:

• Switching to Whole Number loses decimal data
• Using Text increases model size
• Keeping Decimal Number wastes space

4. You notice your Power BI model size is large. You disabled auto date/time but the size did not reduce. What is the likely cause?

medium

A. You need to refresh the data after disabling auto date/time

B. There are many unused columns still in the model

C. Auto date/time disables only for new reports, not existing ones

D. The model size never changes after disabling features

Solution

1. Step 1: Check for unused columns

   Unused columns consume space even if auto date/time is off.

2. Step 2: Remove unused columns

   Deleting these columns will reduce model size effectively.

3. Final Answer:

   There are many unused columns still in the model -> Option B

4. Quick Check:

   Unused columns keep model large despite disabling features [OK]

Hint: Remove unused columns if disabling features doesn't reduce size [OK]

Common Mistakes:

• Assuming refresh fixes size without removing data
• Thinking auto date/time disables only for new reports
• Believing model size never changes

5. You have a large sales dataset with many columns, including detailed customer info not used in reports. You want to reduce model size and improve report speed. Which combination of actions is best?

hard

A. Remove unused customer columns, disable auto date/time, and change numeric columns to appropriate data types

B. Keep all columns, enable auto date/time, and add calculated columns for analysis

C. Remove only unused columns, keep auto date/time enabled, and leave data types as is

D. Disable auto date/time only, keep all columns and data types unchanged

Solution

1. Step 1: Remove unused columns

   Deleting unused customer info reduces unnecessary data in the model.

2. Step 2: Disable auto date/time

   This stops Power BI from adding hidden date tables that increase size.

3. Step 3: Optimize numeric data types

   Changing numeric columns to smaller or fixed decimal types saves space without losing precision.

4. Final Answer:

   Remove unused customer columns, disable auto date/time, and change numeric columns to appropriate data types -> Option A

5. Quick Check:

   Combining removal, disabling, and data type optimization reduces size best [OK]

Hint: Combine removing unused data, disabling features, and data type tuning [OK]

Common Mistakes:

• Keeping all data and enabling auto date/time
• Only removing columns without disabling features
• Ignoring data type optimization