Scatter Plot
Also called: Scatter chart, Scatter graph, Scattergram, XY chart
A scatter plot places one dot per item on two numeric axes, such as ad spend against revenue, so you can see whether the two measures move together and spot outliers.
How to read a scatter plot
Each dot is one item: a store, a customer, a product, a sales rep. Its position comes from two measures, one on each axis. In the chart above, every store is placed by its monthly ad spend (x) and its monthly revenue (y).
Read it in three passes:
- Direction. Do the dots climb from bottom-left to top-right? Then the measures rise together.
- Strength. Are the dots tight around a line, or a loose cloud? A dashed trend line makes this easier to judge.
- Exceptions. Which dots sit far from the rest? Here, one store spends $42k a month but brings in only $190k. That’s the store worth a phone call, and you’d never see it in a total.
Patterns to look for
- Positive: both measures rise together, like spend and revenue.
- Negative: one rises as the other falls, like price and units sold.
- None: no pattern. That’s a useful answer too: the measure you hoped would drive results probably doesn’t.
- Clusters and outliers: separate groups (e.g. small and large stores) or single dots far from everything else.
Common uses
Ordered from most to least common on business dashboards.
1. Quadrant analysis
Draw reference lines at the average of each measure and the chart splits into four groups. Product teams use it for volume vs. margin, sales teams for deal count vs. win rate.
2. Spend vs. return
Marketing spend against revenue or leads, by campaign or by store. The trend line shows the typical return and the outliers show where money isn’t working.
3. Performance vs. effort
Tickets handled against customer satisfaction by agent, or hours logged against output by team. It shows who is fast and good, not just fast.
In Power BI vs. Tableau
| Power BI | Tableau | |
|---|---|---|
| Built-in visual | Scatter chart | A measure on Columns and another on Rows |
| One dot per item | Put the item field (e.g. Store) in Values | Put the item dimension on Detail |
| Trend line | Analytics pane | Analytics pane |
Best practices
- Label the outliers, not every dot. A dozen labels turn into noise; one or two tell the story.
- Say what one dot is in the title or subtitle (“12 stores”, “one dot per customer”).
- Keep both axes honest. Starting an axis above zero is fine on a scatter plot, but say so with clear tick labels.
- Don’t imply cause. Write “stores that spend more tend to earn more”, not “spend drives revenue”.
Mocking up a scatter plot
Scatter plots raise a question stakeholders rarely think about up front: what is one dot? Settle that on a mockup, along with which outliers to label, before anyone builds the measures.
