Choosing the right visual is part art, part science. The wrong chart doesn’t just look bad — it misleads. A bar chart on non-ordinal categories implies an order that doesn’t exist. A pie chart with 15 slices is unreadable. This guide covers the visuals that cover 90% of reporting needs and the rules for choosing between them.
The core visuals
Bar and column charts
When to use: comparing a measure across discrete categories.
- Column chart (vertical bars): 2–12 categories, short labels.
- Bar chart (horizontal bars): 8+ categories or long labels.
// Typical use: revenue by product category
Revenue by Category = SUM(Sales[Revenue]) // plotted against Product[Category]
Common mistakes:
- Starting the y-axis above zero — exaggerates differences.
- Using a column chart when labels overlap — switch to horizontal bar.
- Sorting alphabetically — sort by value so the chart reads as a ranking.
Line charts
When to use: showing trends over a continuous axis (usually time).
- X-axis: date or sequential number.
- Y-axis: a measure that changes over time.
Common mistakes:
- Using a line chart for non-continuous data (e.g., revenue by category). Use a bar chart instead — categories aren’t ordered.
- Too many lines (5+) — the chart becomes spaghetti. Use small multiples or filter to top N.
- Not handling gaps — if data is missing for a period, decide whether to show a gap or interpolate.
Cards and KPIs
When to use: highlighting a single key number.
- Card: just the value, large and prominent.
- KPI visual: value + a target + trend indicator (up/down arrow).
Common mistakes:
- Too many cards on a page — it becomes a dashboard of numbers with no context.
- No comparison — a card showing “$1.2M revenue” is meaningless without “vs $1.0M last month.”
Matrix and table
When to use: showing precise values across two dimensions.
- Table: rows only, for detailed listings.
- Matrix: rows + columns, for cross-tabulations (e.g., revenue by product × region).
Common mistakes:
- Matrix with too many columns — it becomes unreadable. Filter or group.
- Not using conditional formatting — a matrix of raw numbers is hard to scan. Add background color scaling to highlight highs and lows.
Maps
When to use: geographic data (country, state, city, postal code).
- Filled map (choropleth): regions shaded by value. Good for comparisons.
- Bubble map: circles sized by value. Good for showing magnitude at a point.
Common mistakes:
- Using a map when a bar chart would do — if you only have 5 regions, a bar chart is clearer.
- Ambiguous location names — “Springfield” exists in 30+ US states. Use country + state + city for precision.
Scatter plot
When to use: showing the relationship between two measures.
- X-axis: one measure.
- Y-axis: another measure.
- Bubble size (optional): a third measure.
- Play axis (optional): a date field for animated trends over time.
Common mistakes:
- Too many points — the chart becomes a blob. Filter to a subset or use transparency.
- Not labeling outliers — interesting points should be identifiable.
The visual hierarchy
A good report follows a visual hierarchy: top-level summary, supporting detail, drill-down. Users should answer “how are we doing?” in 5 seconds, then dig deeper if needed.
Layer 1: Summary (top of page)
- 2–4 cards showing key metrics (revenue, orders, conversion rate).
- 1 KPI visual per metric with target and trend.
Layer 2: Trend (middle of page)
- A line chart showing the main metric over time (last 12 months).
- A bar chart showing the metric by primary dimension (e.g., by category).
Layer 3: Detail (bottom of page)
- A matrix showing the metric across two dimensions (e.g., category × region).
- A table for users who want exact numbers.
Layer 4: Drill-down (separate page)
- A detailed view accessible via a button or bookmark.
- Full transaction-level data for users who need it.
Choosing between visuals
| Question | Visual |
|---|---|
| How does this change over time? | Line chart |
| How do categories compare? | Bar chart |
| What’s the single most important number? | Card |
| How do two dimensions cross? | Matrix |
| Where are our customers? | Map |
| Is there a correlation? | Scatter plot |
| What’s the composition? | Stacked bar (avoid pie for 5+ slices) |
Common report design mistakes
Mistake 1: Cramming everything on one page. A page with 15 visuals overwhelms users. Aim for 4–7 visuals per page, with clear hierarchy.
Mistake 2: No filtering context. Users need to know what they’re looking at. Add a title, a date range, and any active filters.
Mistake 3: Inconsistent color usage. Use color intentionally — one color for the primary metric, a contrasting color for targets, muted colors for context. Don’t use 10 different colors across 5 visuals.
Mistake 4: Forgetting mobile layout. If the report will be viewed on phones, use Power BI’s mobile layout view to arrange visuals for small screens. The desktop layout doesn’t translate well.
Summary
Most reports need only five visual types: bar, line, card, matrix, and map. Choose based on the question — trends use lines, comparisons use bars, single numbers use cards, cross-tabs use matrices, geography uses maps. Structure reports in layers: summary at the top, trends in the middle, detail at the bottom. Keep pages to 4–7 visuals, use color intentionally, and always provide context with titles and filter indicators.