Creating dataflows in Power BI - Business Scenario Walkthrough

Scenario Mode

👤 Your Role: You are a business intelligence analyst at a retail company.

📋 Request: Your manager wants you to create a reusable dataflow that combines sales and product data to support multiple reports.

📊 Data: You have two data sources: Sales data with columns Date, ProductID, Quantity, and SalesAmount; Product data with columns ProductID, ProductName, and Category.

🎯 Deliverable: Create a dataflow in Power BI that merges sales and product data, cleans it, and makes it available for report building.

Sample Data

Date	ProductID	Quantity	SalesAmount
2024-01-01	101	5	100
2024-01-02	102	3	60
2024-01-03	101	2	40
2024-01-04	103	4	80
2024-01-05	102	1	20

ProductID	ProductName	Category
101	Notebook	Stationery
102	Pen	Stationery
103	Water Bottle	Accessories

1

Step 1: Connect to the Sales data source in Power BI dataflows.

Use Power Query to connect to the Sales data table with columns Date, ProductID, Quantity, SalesAmount.

Expected Result

Sales data is loaded into the dataflow for transformation.

2

Step 2: Connect to the Product data source in Power BI dataflows.

Use Power Query to connect to the Product data table with columns ProductID, ProductName, Category.

Expected Result

Product data is loaded into the dataflow for transformation.

3

Step 3: Merge Sales and Product tables in the dataflow using ProductID as the key.

In Power Query, perform a Merge Queries operation: Sales left join Product on ProductID.

Expected Result

A combined table with sales and product details is created.

4

Step 4: Remove unnecessary columns and rename columns for clarity.

Keep columns Date, ProductID, ProductName, Category, Quantity, SalesAmount. Rename columns if needed.

Expected Result

Cleaned combined table ready for use.

5

Step 5: Save and refresh the dataflow to make it available for reports.

Save the dataflow with a descriptive name like 'SalesProductDataflow' and refresh it.

Expected Result

Dataflow is ready and accessible for report building.

Final Result

---------------------------------------------
| Date       | ProductID | ProductName | Category    | Quantity | SalesAmount |
|------------|-----------|-------------|-------------|----------|-------------|
| 2024-01-01 | 101       | Notebook    | Stationery  | 5        | 100         |
| 2024-01-02 | 102       | Pen         | Stationery  | 3        | 60          |
| 2024-01-03 | 101       | Notebook    | Stationery  | 2        | 40          |
| 2024-01-04 | 103       | Water Bottle| Accessories | 4        | 80          |
| 2024-01-05 | 102       | Pen         | Stationery  | 1        | 20          |
---------------------------------------------

✓The dataflow successfully combines sales and product information.

✓It enables easy access to enriched data for multiple reports.

✓Data is clean and ready for analysis without repeated preparation.

Bonus Challenge

Add a calculated column in the dataflow that computes UnitPrice as SalesAmount divided by Quantity per row.

Show Hint

In Power Query, create a custom column with the formula: [SalesAmount] / [Quantity].

Practice

1. What is the main purpose of creating a dataflow in Power BI?

easy

A. To create visual charts and graphs

B. To prepare and clean data once for use in multiple reports

C. To write DAX formulas for calculations

D. To schedule report email delivery

Solution

1. Step 1: Understand dataflow purpose

   Dataflows are designed to prepare and clean data centrally.

2. Step 2: Compare options

   Creating visuals and writing DAX are report tasks, not dataflow tasks.

3. Final Answer:

   To prepare and clean data once for use in multiple reports -> Option B

4. Quick Check:

   Dataflows centralize data prep = A [OK]

Hint: Dataflows focus on data prep, not visuals or formulas [OK]

Common Mistakes:

• Confusing dataflows with report visuals
• Thinking dataflows create charts
• Assuming dataflows handle email scheduling

2. Which of the following is the correct first step to create a dataflow in Power BI Service?

easy

A. Create a new report and add visuals

B. Open Power BI Desktop and click 'Get Data'

C. Go to workspace > New > Dataflow

D. Write a DAX measure in the report

Solution

1. Step 1: Identify dataflow creation location

   Dataflows are created in Power BI Service under a workspace.

2. Step 2: Verify correct menu path

   The correct path is Workspace > New > Dataflow, not Power BI Desktop or report creation.

3. Final Answer:

   Go to workspace > New > Dataflow -> Option C

4. Quick Check:

   Create dataflow in service workspace = C [OK]

Hint: Dataflows start in Power BI Service workspace, not Desktop [OK]

Common Mistakes:

• Trying to create dataflows in Power BI Desktop
• Confusing dataflows with reports
• Starting with DAX instead of data preparation

3. Consider this scenario: You create a dataflow that connects to an Excel file and applies a filter to keep only rows where Sales > 1000. What will happen when you refresh the dataflow?

medium

A. The dataflow will load all rows from Excel without filtering

B. The dataflow will refresh only if the Excel file is open

C. The dataflow will fail because filters are not allowed

D. The dataflow will load only rows where Sales > 1000

Solution

1. Step 1: Understand dataflow refresh behavior

   Dataflows apply all transformations including filters during refresh.

2. Step 2: Apply filter effect

   Only rows meeting Sales > 1000 condition are loaded into the dataflow.

3. Final Answer:

   The dataflow will load only rows where Sales > 1000 -> Option D

4. Quick Check:

   Filters apply on refresh = B [OK]

Hint: Filters in dataflows apply automatically on refresh [OK]

Common Mistakes:

• Thinking filters don't apply in dataflows
• Assuming dataflows need source files open
• Believing dataflows load all data regardless of filters

4. You created a dataflow but it fails to refresh with an error saying "Invalid credentials". What is the most likely cause and fix?

medium

A. The data source credentials are missing or expired; update credentials in dataflow settings

B. The dataflow name is invalid; rename the dataflow

C. The Power BI Service is down; wait and retry later

D. The dataflow has too many columns; reduce columns

Solution

1. Step 1: Analyze error message

   "Invalid credentials" means Power BI cannot access the data source due to authentication issues.

2. Step 2: Identify fix

   Updating or re-entering credentials in dataflow settings resolves this issue.

3. Final Answer:

   The data source credentials are missing or expired; update credentials in dataflow settings -> Option A

4. Quick Check:

   Invalid credentials = update data source login [OK]

Hint: Check and update data source credentials when refresh fails [OK]

Common Mistakes:

• Ignoring credential errors and blaming dataflow name
• Assuming service outage without checking credentials
• Thinking column count causes credential errors

5. You want to create a dataflow that combines data from two different SQL databases and refreshes daily. Which approach is best to ensure efficient and maintainable dataflows?

hard

A. Create two separate dataflows for each database, then create a third dataflow to merge their outputs

B. Create one dataflow that connects to both databases and merges data in a single query

C. Create one dataflow per database and merge data only in Power BI Desktop reports

D. Create multiple reports instead of dataflows to handle merging

Solution

1. Step 1: Consider modular design

   Separating dataflows by source improves maintainability and reusability.

2. Step 2: Combine dataflows for merging

   Creating a third dataflow to merge outputs centralizes transformation and refresh logic.

3. Final Answer:

   Create two separate dataflows for each database, then create a third dataflow to merge their outputs -> Option A

4. Quick Check:

   Modular dataflows + merge = A [OK]

Hint: Split sources, then merge in a separate dataflow for clarity [OK]

Common Mistakes:

• Merging different sources in one complex dataflow
• Merging only in reports causing repeated work
• Avoiding modular dataflows for easier updates