7.1 Visual Types and Selection

Quick Answer: Choose visuals by what you are analyzing: bar/column for comparison, line for trends, pie/donut for composition, scatter for relationships, cards for KPIs, and tables/matrices for detail. The PL-300 tests selecting the most appropriate visual for a given business scenario.

Choosing the Right Visual

Core Visual Types

Bar and column charts compare values across categories. Use horizontal bars when category labels are long (product names) and vertical columns when categories are few with short labels. Stacked variants add composition; 100% stacked shows proportions.

Line and area charts show trends over continuous time. Line-and-clustered-column combos display two metrics (for example revenue bars with a margin line) on dual axes.

Pie and donut charts show parts of a whole. Limit to about 5-7 slices; for many categories prefer a treemap or stacked bar.

Tables and matrices display detail. A matrix adds row and column hierarchies with expand/collapse, subtotals and grand totals, stepped layout, and conditional formatting, the pivot-table style.

Cards surface a single metric: Card for one big number, Multi-row card for several, and KPI for value plus trend and target.

Maps require geographic data: Map (bubbles), Filled map (colored regions), Shape map (custom boundaries), and Azure map/ArcGIS for advanced spatial analysis. Set the relevant Data Category (City, State, Country, Latitude/Longitude) so geocoding works.

Visual Calculations (DAX in Visuals)

Visual calculations let report authors write DAX directly on a visual, scoped to that visual's grid, without adding model measures. They reached general availability in the May 2026 Power BI update and are available across Power BI Pro, Premium Per User, Premium capacity, and Fabric F-SKU tiers.

Running Total = RUNNINGSUM([Revenue])
Vs Previous = [Revenue] - PREVIOUS([Revenue])
3-Month Avg = MOVINGAVERAGE([Revenue], 3)

Key visual-calculation functions include RUNNINGSUM, MOVINGAVERAGE, PREVIOUS/NEXT, FIRST/LAST, COLLAPSE/EXPAND, RANGE, and WINDOW. Because they operate on the visual's already-aggregated rows, they make running totals and percent-of-parent far simpler than the equivalent model DAX.

Paginated Reports

Paginated reports are pixel-perfect, print-ready reports for fixed layouts and full-row detail.

Choose paginated for invoices, statements, fixed page dimensions, full-row output, or print/PDF distribution.

On the Exam

The PL-300 tests selecting the right visual for a scenario, tables vs. matrices, when paginated reports apply, visual calculations vs. model measures, and map data-category requirements.

Match the Visual to the Question, Not the Data

The exam's visual-selection items are really about analytical intent. Decide first what the user is trying to learn, compare, trend, compose, relate, distribute, or spotlight a single number, and the right visual follows. Comparisons across categories call for bar or column charts; trends over continuous time call for line charts; part-to-whole at a glance calls for pie, donut, or stacked bars (kept to a few segments); two-variable relationships call for scatter; and a single headline metric calls for a card or KPI. Reaching for a fancy custom visual when a clustered column answers the question is a classic wrong answer.

Tables vs. Matrices, and When Detail Beats Charts

A table is a flat list of rows and is the right choice when users genuinely need to read or export individual records. A matrix pivots data with row and column hierarchies, subtotals, grand totals, and stepped layout, the choice when users want cross-tabulated analysis like product-by-region. Both support conditional formatting (data bars, color scales, icons) that turns dense numbers into a scannable grid. When a report must show every transaction, fit a printed page, and include headers/footers, neither belongs on an interactive canvas, that is the paginated report scenario, authored in Power BI Report Builder.

Visual Calculations Change the Authoring Calculus

Now that visual calculations reached general availability in May 2026 across Pro, PPU, Premium, and Fabric tiers, report authors can express visual-scoped logic, running totals, moving averages, percent-of-parent, period-over-period deltas, directly on the visual with functions like RUNNINGSUM, MOVINGAVERAGE, and PREVIOUS, without cluttering the semantic model with single-use measures.

The exam-relevant judgment is scope: if the calculation only makes sense within one visual's grid and is not reused, a visual calculation is cleaner; if many visuals or reports need it, author a model measure so the logic stays centralized and governed.