Tableau Scatter Plots: Correlation and Outliers for Analysts
Build Tableau scatter plots for correlation and outliers: map Sales vs Profit, add Detail, Size, Color and fix overplotting.
When a stakeholder asks "do higher sales mean higher profit per customer?" a table of averages hides the answer. A scatter shows the slope, the spread and the outliers that averages erase. Built from script 05, this guide walks the build, encoding and clean-up so your scatter is insight, not ink.
Anchor in the Data Analyst Roadmap, encode via Tableau Marks Card, and model correctly via Tableau Relationships vs Joins.
When should you use a scatter plot?
Correlation and outlier detection for two continuous measures Tableau Help: Scatter and Cleveland. Scatter plots answer:
- Is Sales positively correlated with Profit?
- Which customers are outliers (high Sales, negative Profit)?
- Is the relationship linear or clustered?
Bars aggregate; scatters preserve individual points. Choose scatter when the entity (Customer ID, Product) matters, not just the category total.
How do you build a basic scatter?
Columns + Rows + Detail Tableau Help: Scatter. Steps from script at 0:27:
- Drag Sales (Measure) to Columns.
- Drag Profit (Measure) to Rows.
- Drag Customer ID (Dimension) to Detail on the Marks Card.
Tableau now plots one circle per customer at (Sales, Profit). Without Detail, Tableau would aggregate to a single point (SUM). Detail sets the granularity.
// Build
// Columns: SUM([Sales]) Rows: SUM([Profit]) Detail: [Customer ID]Test: remove Customer ID from Detail and the scatter collapses to one mark — proof that Detail defines entity level.
How do you encode more dimensions?
Add channels via the Marks Card Tableau Help: Marks.
- Size: Map
COUNT(Orders)or Quantity to Size — larger circles mean more orders. - Color + Shape: Add Country or Segment to Color and Shape — Tableau assigns hue and symbol per value.
- Filled Circles: On Marks Card, switch Shape to Filled Circle for density readability.
Example:
// Advanced encoding
// Size: COUNT([Order ID]) Color: [Country] Shape: [Country]Four dimensions at once: X=Sales, Y=Profit, Size=orders, Color/Shape=Country. Keep to ≤4 encodings; more harms perception Cleveland.
| Feature / Criteria |
|---|
How do you fix overplotting?
Opacity, Border and filtering Tableau Help: Marks. Dense scatters at 5k points hide density:
- Opacity: Click Color > Opacity > 60-70%. Overlapping points darken, revealing density.
- Border: Under Color > Effects, add a thin dark Border so individual points separate.
- Filter or aggregate: If still crowded, filter to a Segment or add a dashboard Filter Action (see Tableau Dashboards).
Gotcha: Overplotting Lies
You present a 10k-point scatter at 100% opacity with no border. Stakeholders see a solid blob and miss the cluster of high-sales negative-profit outliers. Lower opacity to 60% and add a 1px border — the cluster pops, and the insight lands. Always test prints at 70% zoom.
Quick reference
| Build Step | Action |
|---|---|
| Base scatter | Sales to Columns, Profit to Rows, Customer ID to Detail |
| Add size | COUNT(Order ID) to Size |
| Add segment | Country to Color and Shape |
| Clean overplot | Color > Opacity 65%, Border thin, Filled Circles |
| Next | Move to Tableau Calculated Fields to derive Cost |
Practice by reproducing the script's scatter, then add Size and Color and screenshot before/after overplotting fix. Publish the clean version to your portfolio.
Turn Scatters into Portfolio Stories
Build one scatter that finds an outlier, then add a caption explaining the business implication on your Data Analyst track.
Explore the Data Analyst TrackFrequently Asked Questions
When should you use a scatter plot in Tableau?
Use a scatter plot when analysing correlation between two continuous measures and spotting outliers. Each point is an entity like a customer, plotted by two measures such as Sales and Profit.
How do you build a scatter plot in Tableau?
Drag one measure to Columns (Sales) and one to Rows (Profit), then drag a dimension like Customer ID to Detail. Tableau creates one mark per customer at the intersection of the two measures.
How do you add a third and fourth dimension to a scatter?
Map a measure to Size for a third channel, and a dimension to Color and Shape for categorical segmentation. Convert marks to Filled Circles for readability.
How do you fix overplotting in dense scatters?
Reduce Color opacity, add a Border under Color effects, and consider aggregating or filtering to a relevant segment. Overplotting hides density without these fixes.
What insights does a scatter add over a bar?
Bars compare aggregates per category, while scatters show distribution, correlation direction, and individual outliers — essential for customer-level diagnostics.
Written by
Founder at Topfolio with 6+ years in data & analytics across JPMC, Ultrahuman, and high-growth startups. Sat on hiring panels, reviewed 500+ resumes, and writes practical SQL & data guides.
Related Articles
Dimensions vs Measures in Tableau: Blue vs Green Pills Guide
Understand Dimensions vs Measures and Discrete vs Continuous in Tableau. Learn why blue pills create headers, green pills create axes, and how to convert them.
Tableau Table Calculations: Running Total, Difference and Moving Average
Use Tableau Quick Table Calculations: Running Total, Difference and 14-day Moving Average with continuous dates.
Tableau Complete Guide: Master Business Intelligence and Dashboards (2026)
Master Tableau for business intelligence: data connections, visual analytics, calculated fields, LOD expressions, table calculations, and interactive dashboards.