Pick a bar chart for categorical comparisons, a line chart for time‑series trends, a scatter plot for relationship analysis, and a treemap for hierarchical proportion.
1. Identify the data grain – if the dimension is discrete (product, region) use bars; if it’s continuous (date, timestamp) use lines.
2. Count categories or series – bars become cluttered >20 items, lines >12 series lose color distinction, scatter >500 points should enable density shading or clustering, treemap >100 leaf nodes slows rendering.
3. Define the analytical goal – compare values → bar, show trend → line, discover correlation → scatter, illustrate share of a hierarchy → treemap.
4. Validate with thresholds – in Power BI set VisualLevelFilters:
```dax
// Example: limit bar categories to top 20 by sales
TopCategories =
TOPN(20, VALUES(Sales[Region]), [Total Sales], DESC)
```
```json
{
"visual": "BarChart",
"maxCategories": 20,
"showLegend": true
}
```
5. Test interaction – enable cross‑filtering; for scatter, add a trend line via Analytics > Add > Line and set ConfidenceLevel = 95.
Quick comparison
| Visual | Best for | Max categories/points | Typical KPI |
|---|---|---|---|
| Bar | Discrete categories | ≤20 | Revenue by region |
| Line | Continuous time | ≤12 series | Monthly active users |
| Scatter | Bivariate numeric | ≤500 (use clustering) | Sales vs ad spend |
| Treemap | Nested hierarchy | ≤100 leaf nodes | Product line contribution |
Use Power BI’s Visualization > Format > Data colors to assign corporate palette, and set Title > Text to include the metric name and period for executive clarity.