Getting started with Power BI

Power BI Desktop is a free Windows application for turning raw files and databases into interactive reports. The workflow is always the same: get data, clean it, model it, visualise it, publish it.

Install and look around

Download Power BI Desktop from the Microsoft Store (it updates itself every month). On the left edge you will find three views:

View	Purpose
Report	The canvas where you build charts, cards and slicers
Table (Data)	Inspect the loaded rows, add columns and measures
Model	See tables and draw relationships between them

Step 1: Get data

Home, Get data connects to Excel, CSV, JSON, SQL Server, MySQL, PostgreSQL, SharePoint, a web page and dozens more. Pick a monthly sales export to begin with. In the Navigator window choose the sheet or table, then click Transform data rather than Load, so the data passes through Power Query first.

Step 2: Clean in Power Query

The Power Query editor records each click as a step you can rename, reorder or delete under Applied Steps. For a typical branch sales export you would:

• promote the first row to headers,
• set Order Date to Date, Amount to Decimal Number, Qty to Whole Number,
• remove blank rows and unused columns,
• trim and clean the City text,
• filter out the "TOTAL" summary rows at the bottom.

= Table.TransformColumnTypes(
      Source,
      {{"Order Date", type date}, {"Amount", type number}, {"Qty", Int64.Type}}
  )

Click Close & Apply to load the cleaned table into the model.

Step 3: First visuals

On the Report view, drag a field onto the canvas and Power BI guesses a visual; then change it from the Visualizations pane. A first page for a retail chain might have:

• a Card showing Sum of Amount as total sales,
• a Clustered column chart with City on the axis and Amount as values,
• a Line chart with Order Date on the axis to show the monthly trend,
• a Slicer for Region so a manager can filter everything at once.

Selecting a bar filters the other visuals automatically. That cross-filtering is what makes a Power BI report different from a static Excel chart.

Step 4: A first measure

Fields dropped into a visual are aggregated implicitly, but named measures are reusable and easier to maintain.

Total Sales = SUM(Sales[Amount])
Order Count = COUNTROWS(Sales)
Average Order Value = DIVIDE([Total Sales], [Order Count])

Create them with Modeling, New measure, and set the format (for example a rupee currency format with zero decimal places) once, in the Measure tools ribbon.

Step 5: Publish

Save the .pbix file, then Home, Publish and choose a workspace in the Power BI service. Colleagues open the report in a browser, and you can schedule a daily refresh so the numbers keep updating.

Key takeaways

• Always go through Transform data; fixing types later is painful.
• Power BI files hold data, model and report together in one .pbix.
• Build measures instead of relying on implicit aggregations.
• Publishing, not emailing the file, is how a report gets shared.