Skip to content

05 · Building Your First Visuals

A visual in Power BI is a question sent to the model: "group by these fields, compute these numbers." The chart type is only the way the answer is drawn. Getting the question right — which fields go where, and how numbers aggregate — matters more than picking a pretty chart.

The Report view

  • Canvas in the middle, with page tabs at the bottom.
  • Visualizations pane: a gallery of visual types and, below it, the field wells (called "buckets" in some docs) for the selected visual — for example X-axis, Y-axis, Legend, Tooltips.
  • Data pane (formerly "Fields"): your tables and columns.
  • Format tab inside the Visualizations pane (paint-roller icon) for styling.

Pane names and icons have been reshuffled more than once across monthly releases; if a pane isn't where this lesson says, look under the View ribbon to toggle it.

Step by step: revenue by region

Assuming the Sales table with the Revenue column from lesson 03:

  1. Click an empty spot on the canvas, then click Clustered column chart in the Visualizations gallery.
  2. Drag Region to X-axis and Revenue to Y-axis.
  3. The Y-axis well shows "Sum of Revenue" — Power BI applied implicit aggregation.
  4. Hover a column. Expected values (from the sample):
Region Sum of Revenue
South 700
North 445
East 280
  1. Columns are sorted by value descending by default. Use the More options (…) → Sort axis menu on the visual to switch to alphabetical, and back.

Picking the right visual

Question Good default Avoid
Compare categories Bar/column chart Pie with more than ~5 slices
Change over time Line chart (dates on X) Column chart with 60 months
One headline number Card (or the newer card visual) A gauge unless there is a real target
Exact values to look up Table or matrix Charts with data labels on every point
Part-to-whole, few parts Stacked bar, or 100% stacked 3-D anything
Two measures' relationship Scatter chart Dual-axis charts with unrelated scales

Add these three to the page for practice:

  • Card: Revenue → shows 1.43K by default (display units "Auto"). Set Display units to None in Format → Callout value to show 1,425.
  • Line chart: OrderDate on X-axis, Revenue on Y. Power BI may create a date hierarchy (Year → Quarter → Month → Day). Use the drill arrows at the top of the visual to expand to month level. Expected: Jan 605, Feb 380, Mar 440.
  • Table: Category, Units, Revenue. Expected: Camping 6 units / 720, Apparel 6 / 480, Accessories 9 / 225, total 21 / 1,425.

The implicit-aggregation trap

Drag UnitPrice into the table next to Category. Power BI shows "Sum of UnitPrice": Camping = 120 + 120 + 120 = 360. That number is meaningless — you never sell a tent at 360. Change the aggregation (dropdown on the field in the well → Average) and you get 120, but even that is an unweighted average of rows, not revenue ÷ units.

The fix is not to rely on implicit aggregation for anything that matters. In lesson 07 you will write explicit measures such as Avg Selling Price = DIVIDE([Total Revenue], [Total Units]). You can also stop the model from summing a column automatically: select UnitPrice in the Data pane and set Column tools → Summarization → Don't summarize. Do the same for OrderID, which is a number but should never be added up.

How It Actually Works

When a visual renders, Power BI generates a DAX query and sends it to the local Analysis Services engine. You can see it: View → Performance Analyzer → Start recording → Refresh visuals, expand the column chart's entry, and click Copy query. Stripped of some wrapping, it looks like this:

EVALUATE
TOPN(
    1002,
    SUMMARIZECOLUMNS(
        'Sales'[Region],
        "SumRevenue", CALCULATE(SUM('Sales'[Revenue]))
    ),
    [SumRevenue], 0,
    'Sales'[Region], 1
)
  • SUMMARIZECOLUMNS groups by the fields in the axis/legend wells and computes each value in the value wells, for every group that has data.
  • The implicit "Sum of Revenue" becomes SUM('Sales'[Revenue]). It is a hidden measure generated on the fly.
  • TOPN(1002, …) limits the rows returned. Visuals use data reduction limits so a chart with millions of categories doesn't try to draw them all; tables fetch more rows as you scroll.

The engine returns three rows for three regions, and the visual draws three columns. Totals in tables and matrices come from separate evaluation at the "grand total" level — not from adding up the displayed rows — which is why a measure can legitimately show a total that isn't the sum of its rows (an average, a distinct count). Lesson 07 builds on that.

Common mistakes

  • Summing columns that aren't additive: prices, percentages, IDs, ages, balances.
  • Too many categories in a pie or donut. If you can't read the labels, it's the wrong visual.
  • Using the automatic date hierarchy without knowing it. Power BI creates a hidden date table per date column when Auto date/time is on (File → Options → Data Load). Fine for learning; Level 2 replaces it with a proper date table.
  • Formatting before the numbers are right. Check totals against a known answer first.

Exercise

  1. Build the four visuals described and confirm every number against the tables above.
  2. Use Performance Analyzer to copy the DAX query for your table visual. Identify the group-by columns and the aggregations in it.
  3. Set OrderID and UnitPrice to Don't summarize, then try to drag UnitPrice into a card. Describe what the card shows and why.