Power BI: turn data into decisions
Build your first useful Power BI report: prepare a small dataset, connect a simple model, create clear measures, and explain what the visuals mean.
Learn it here, one small step at a time.
4 guided lessons · About 2h 30m including practice · No subscription
- Prepare and load data for a specific question
- Build a simple fact-and-dimension model
- Create readable visuals and explain their limits
Recommended first: Excel: from first cell to confident analysis, SQL foundations: ask better questions. These foundations are suggestions, not a requirement to begin.
Start with a business question
Power BI helps turn data into a report someone can use. Start by asking one question, such as “Which region sold the most this month?” Identify the audience, decision, date range, and unit before choosing a chart. A chart is not useful just because it looks polished.
Power BI Desktop can import a CSV or workbook. In the preview, check that headers are correct, dates are dates, and amounts are numeric. Keep a note of the source and its refresh date.
| Decision | Useful field | Check |
|---|---|---|
| Compare areas | Region | Are region names consistent? |
| Track money | Amount | Is this currency and numeric? |
| Choose a time period | Date | Are dates real dates? |
Write one question your report should answer. List the data columns needed to answer it and one column that is not needed.
Reveal a suggested answer
The question determines which data and comparison matter, preventing a decorative report with no clear purpose.
Optional further reading
Core lesson and practice are on this page. For more detail, visit the official free reference ↗.
Prepare data and model relationships
Power Query opens before data is loaded. Remove only errors or duplicates you understand; set each column to the right type and keep the original source untouched. A useful model often has a sales fact table (many transactions) connected to small lookup tables, such as Date, Product, or Region.
A one-to-many relationship means one lookup row can relate to many sales. Use a unique key on the “one” side. If totals suddenly double, inspect the relationship and whether the lookup key is actually unique.
| Table | Example row | Purpose |
|---|---|---|
| Sales (fact) | one sale, amount $12 | Records events and numbers |
| Product (lookup) | Notebook, category Paper | Describes a product once |
| Date (lookup) | May 2, 2026 | Makes time filters consistent |
Draw a line from `Product[ProductID]` to `Sales[ProductID]`. Which side should contain each product only once?
Reveal a suggested answer
A lookup table keeps descriptive values unique, so relationships filter facts predictably and avoid double-counted results.
Optional further reading
Core lesson and practice are on this page. For more detail, visit the official free reference ↗.
Create a measure and choose a chart
A measure calculates a result from the current filters. A basic DAX measure is Total Sales = SUM(Sales[Amount]). Use a measure for totals instead of adding up values in a chart by hand. Put Region on the category axis and Total Sales in Values for a bar chart.
Use a line chart for change over time, not for unordered categories. Add a descriptive title and units. Keep colors consistent and remove anything that does not help answer the report question.
| Question | Useful visual | Reason |
|---|---|---|
| Which region is larger? | Bar chart | Categories are easy to compare |
| How did sales change each month? | Line chart | Shows ordered time movement |
| What is the total? | Card | Highlights one key number |
Create the `Total Sales` measure. Make a bar chart of sales by region and label its currency and date range.
Reveal a suggested answer
A measure is recalculated as filters change, so the displayed total can reflect the selected region or date.
Optional further reading
Core lesson and practice are on this page. For more detail, visit the official free reference ↗.
Read, check, and share the report
A good report states what its numbers include. Add a subtitle or note with the time range, currency, and source. Test a slicer: does it change every visual that should respond? Compare one region’s visual total with a simple total in the source data.
Before sharing, check text size, contrast, keyboard order, and whether the chart can be understood without color alone. Avoid exposing personal or confidential details in a public report.
| Review | Pass when… |
|---|---|
| Accuracy | A sample total matches the source |
| Clarity | A reader can state the main finding |
| Accessibility | Labels are readable and color is not the only cue |
| Privacy | The report contains only approved data |
Write a one-sentence finding using the chart. Include its date range and avoid claiming the chart proves why a region performed differently.
Reveal a suggested answer
A chart can show that values changed; it does not by itself explain why they changed.
Optional further reading
Core lesson and practice are on this page. For more detail, visit the official free reference ↗.
Your final knowledge check.
Score 80% or more to pass. Sign in to save your best result.