Scatter Plots

Build a scatter plot in Tableau to compare two continuous measures, encode a third and fourth dimension using Size, Color, and Shape, switch marks to Filled Circles, and fix overplotting with Opacity and Border adjustments.

4 minutesVideo LessonPDF notes

🎯 Free Guest Mode: You are learning for free. Sign in to save your completion progress and quiz answers.

Loading progress...

Key moments

Scatter Plot Core Purpose — The instructor outlines scatter plots as tools for identifying correlations and outliers between two continuous measures.
Building the Base Axis — Sales and Profit are dragged to Columns and Rows to create continuous horizontal and vertical axes.
Disaggregating via Detail — Customer ID is dropped onto the Detail shelf to unbundle grand totals into individual customer marks.
Adding Size Encoding — Order count is mapped onto the Size shelf so circle diameters reflect customer purchase volumes.
Color and Shape Mapping — Country is placed on Color and Shape to categorize marks using hues and distinct glyphs.
Formatting Density and Borders — Marks are changed to filled circles with reduced opacity and gray borders to clarify overlapping data.

PDF notes

Bottom line: You will construct scatter plots in Tableau to evaluate correlations and detect outliers between continuous measures. You will control mark granularity using the Detail shelf and encode additional analytical dimensions using Color, Size, Shape, opacity, and borders.

This lesson on Scatter Plots is hands-on and example-driven. You will construct scatter plots in Tableau to evaluate correlations and detect outliers between continuous measures. You will control mark granularity using the Detail shelf and encode additional analytical dimensions using Color, Size, Shape, opacity, and borders.

What You'll Be Able To Do

Plot two continuous measures on Columns and Rows to create a two-dimensional coordinate space.
Disaggregate aggregated marks across a dimension by placing a granular identifier on Detail.
Encode a third numerical measure onto mark Size to represent volume or quantity per data point.
Differentiate categories across data points using the Color and Shape marks cards.
Format overlapping data points by adjusting Mark type to Circle, decreasing opacity, and applying borders.

Detailed Concept Walkthrough

1. Two-Dimensional Coordinate Setup

Scatter plots display the relationship between two continuous measures by mapping one to the horizontal axis and the other to the vertical axis. This reveals linear or non-linear correlation patterns and isolated outlier values.

Mechanism: Dragging a continuous measure like Sales to Columns creates the horizontal X-axis, while dragging Profit to Rows builds the vertical Y-axis. Tableau automatically aggregates these fields to SUM(Sales) and SUM(Profit) by default.
Under the Hood: In its initial state with only two aggregated measures, Tableau renders a single mark representing the grand total coordinates across the entire data source.
Best Practice: Always verify that both axes represent numeric, continuous fields (green pills) rather than discrete categories to ensure a continuous Cartesian space.

// Tableau Shelf Placement for Base Scatter Coordinates:
Columns: [Sales]      // Continuous measure (X-axis)
Rows:    [Profit]     // Continuous measure (Y-axis)
Marks:   Automatic    // Renders single point at SUM(Sales), SUM(Profit)

Key Takeaway: Placing two continuous measures on Columns and Rows sets up the 2D Cartesian plane for correlation analysis.

2. Granularity via Detail Shelf

The Detail shelf determines the level of aggregation (granularity) of the visualization without altering the visual mark type or Cartesian axes. It fractures aggregated marks into individual dimension members.

Mechanism: Placing a discrete dimension such as Customer ID on the Detail card breaks the single aggregated coordinate into one distinct mark for every customer in the dataset.
Under the Hood: Tableau generates a GROUP BY query including the dimension on Detail, calculating SUM(Sales) and SUM(Profit) for each distinct Customer ID record.
Execution Flow: Once unbundled, data points scatter across the plane, allowing immediate visual detection of outliers (isolated points) and general correlation trajectories.

// Adding granularity to scatter plot:
Columns: [Sales]
Rows:    [Profit]
Marks Card -> Detail: [Customer ID] // Disaggregates data to customer-level marks

Key Takeaway: The Detail shelf controls mark disaggregation without adding axes or visual encodings.

3. Multi-Attribute Visual Encoding

Marks cards allow multi-dimensional analysis on a single scatter plot by binding additional measures or dimensions to Size, Color, and Shape.

Mechanism: Dragging a continuous measure like order count (CNT(Orders)) to Size scales individual marks based on transaction volume. Dragging a discrete dimension like Country to Color and Shape assigns distinct hues and glyphs per country.
Under the Hood: Tableau applies a secondary legend mapping for each encoding, mapping quantitative values to pixel radii and categorical values to palette indexes.
Best Practice: When mapping categories to Shape, avoid exceeding 7-10 distinct values to prevent visual clutter and indistinguishable symbols.

// Multi-attribute encoding configuration:
Marks Card -> Size:  [CNT(Orders)] // Sizes circles by order volume
Marks Card -> Color: [Country]     // Colors circles by categorical country
Marks Card -> Shape: [Country]     // Assigns distinct shapes per country

Key Takeaway: Layering measures on Size and dimensions on Color or Shape embeds multi-variable insights into a single 2D plane.

4. Overplotting and Mark Styling

High-density scatter plots suffer from overplotting where overlapping marks obscure point density. Changing mark types, lowering opacity, and adding borders clarifies mark distribution.

Mechanism: Changing the Marks dropdown from Shape/Automatic to Circle produces solid filled marks. Clicking the Color card exposes both the Opacity slider and the Border styling options.
Execution Flow: Lowering opacity causes overlapping circles to visually blend into darker, more saturated regions, instantly revealing local data point density.
Best Practice: Apply a subtle gray or dark border around filled circles so adjacent and overlapping transparent marks retain distinct geometric perimeters.

// Styling marks for high-density readability:
Marks Dropdown -> Select 'Circle'
Marks Card -> Color -> Opacity: 60%
Marks Card -> Color -> Effects -> Border: Select 'Gray'

Key Takeaway: Filled circles combined with reduced opacity and perimeter borders resolve mark occlusion in dense scatter plots.

Topics Covered in Scatter Plots

Scatter Plot Core Purpose (0:00 - 0:25) — The instructor outlines scatter plots as tools for identifying correlations and outliers between two continuous measures.
Building the Base Axis (0:25 - 0:52) — Sales and Profit are dragged to Columns and Rows to create continuous horizontal and vertical axes.
Disaggregating via Detail (0:52 - 1:14) — Customer ID is dropped onto the Detail shelf to unbundle grand totals into individual customer marks.
Adding Size Encoding (1:14 - 1:35) — Order count is mapped onto the Size shelf so circle diameters reflect customer purchase volumes.
Color and Shape Mapping (1:35 - 1:55) — Country is placed on Color and Shape to categorize marks using hues and distinct glyphs.
Formatting Density and Borders (1:55 - 2:35) — Marks are changed to filled circles with reduced opacity and gray borders to clarify overlapping data.

Reference Cheat Sheet

Columns / Rows Shelves — Establishes X and Y coordinates for continuous correlation
```
Columns: [Sales] | Rows: [Profit]
```
Marks -> Detail — Splits aggregated measures into distinct marks per dimension member
```
Marks -> Detail: [Customer ID]
```
Marks -> Size — Scales mark radius proportionally to a quantitative measure
```
Marks -> Size: [CNT(Orders)]
```
Marks -> Color — Applies categorical hues or continuous gradients to marks
```
Marks -> Color: [Country]
```
Marks -> Shape — Assigns distinct geometric glyphs to categorical dimension members
```
Marks -> Shape: [Country]
```
Marks Dropdown -> Circle — Converts default marks into uniform filled circles
```
Marks: Circle
```
Color -> Opacity & Border — Reduces fill opacity and adds contrasting mark borders
```
Color -> Opacity: 65% | Border: #808080
```

Comparison Table

Shelf / Property	Data Type Accepted	Visual / Structural Impact
Columns & Rows	Continuous Measures	Defines 2D Cartesian axes coordinates
Detail Shelf	Discrete Dimensions	Disaggregates marks without visual styling
Size Shelf	Continuous Measures	Varies circle diameter by value
Color Shelf	Measures or Dimensions	Applies palette gradients or categories
Shape Shelf	Discrete Dimensions	Replaces circle with custom glyphs
Color -> Opacity	Slider (0% - 100%)	Shows mark density in overlaps

Common Pitfalls

Mistake: Leaving continuous measures on Columns and Rows without adding a dimension to Detail. Avoid: Add a granular identifier like Customer ID to Detail to disaggregate the single summary mark.
Mistake: Retaining 100% opacity on dense scatter plots with overlapping points. Avoid: Lower opacity in the Color menu so clustered marks darken and reveal point density.
Mistake: Using unfilled default shapes for dense data points. Avoid: Switch the Marks dropdown to Circle and enable a perimeter border under Color effects.

FAQs

Why does Tableau show only one point when I add Sales to Columns and Profit to Rows? Tableau aggregates all rows into grand totals by default (SUM). Adding a dimension to the Detail shelf disaggregates the view into individual marks.
What is the primary analytical purpose of a scatter plot? Scatter plots identify correlations (positive, negative, non-linear) between two continuous measures and locate isolated outlier data points.
How do opacity and borders help with overlapping marks? Lower opacity creates darker shading where points overlap to show density, while borders keep individual mark boundaries discernible.
Can I encode both Color and Shape using the same dimension? Yes, dropping a dimension onto Color and Shape simultaneously provides redundant encoding, making categories easier to identify across the chart.

Common pitfalls

Mistake: Leaving continuous measures on Columns and Rows without adding a dimension to Detail. Avoid: Add a granular identifier like Customer ID to Detail to disaggregate the single summary mark.
Mistake: Retaining 100% opacity on dense scatter plots with overlapping points. Avoid: Lower opacity in the Color menu so clustered marks darken and reveal point density.
Mistake: Using unfilled default shapes for dense data points. Avoid: Switch the Marks dropdown to Circle and enable a perimeter border under Color effects.

Okay, so now we're going to learn how to create the scatter plots in Tableau. Scatter plots are one of the fundamental charts in order to understand the relationship between two continuous measures. So that means the main task of the scatter plots is to find correlations between two continuous fields. And as well another task of the scatter plot is to find the outliers inside your data. So let's go now and create a very basic scatter plots in Tableau, and as I said, we need two measures in order to do that. So our two measures going to be the Sales and the Profit. So let's get the Sales to the Columns and as well the Profit to the Rows. So with that, we got our two axes, and it going to represent a two-dimensional graph. So now what is missing is, of course, our data, the data points. So here we're going to go with the Customer ID. So let's take the Customer ID and now we're going to go and put it to the Detail. And here is the power of Tableau compared to any other tools, where Tableau going to go and plot all data points that we have inside our data without any restrictions. So with that, we can see the correlation between the Sales and the Profit and as well to find the outliers, for example, those points that we have it as standalone. All right, so with that, we have created the very basic scatter plots in Tableau. All right, so next we're going to go and add more stuff to the design of the scatter plots, where we're going to change the colors, the size, add circles, and so on. So now we're going to go and change the size of each data points, but it's going to depend on a third measure, the count of orders. So now let's go to the Orders count and drag and drop it to the Size. Each customer is going to has different size and that's going to depend on how many orders did this customers place. So this is one thing that we can add to our scatter plot. Another thing, we can add coloring. So here we have different ways on how to add colorings. Either we're going to add a dimension or we can make a cluster. So now, for example, let's go and get the dimension Country and place it on the Color. And here in the data points, we can add as well different shapes in our visuals. So currently we have the circle for everything. We can take the Country, drag and drop it to the Shape, and now we can see in the scatter plot, not only that the countries has different colors, but they have as well different shapes. But what we usually see in the scatter plots is that each data point is going to be represented as a filled circle. So that means we're going to go and change the visual. Let's go to the Marks over here, and then change it from Shape to Circle. So now as you can see, we have everything as a filled circle, but we are not there yet. Let's go and make the size a little bit bigger. And now what do we have over here? We have a lot of points, and what we usually do, we go and reduce the opacity of the colors. So let's go to the Color over here and let's just reduce it. And with that, you can see very nicely, for example, those two points, there is like overlapping between them. One more thing that we can add to those circles, we can have like a line border for each circle. So in order to do that, we're going to go again to the Color, and here we have an effect called Border. So instead of Automatic, let's have something like this color, the gray. So with that, you can see we have a very nice border for each data points. All right, so those are some different options on how to customize the scatter plots.

Frequently asked questions

Why does Tableau show only one point when I add Sales to Columns and Profit to Rows?

Tableau aggregates all rows into grand totals by default (`SUM`). Adding a dimension to the Detail shelf disaggregates the view into individual marks.

What is the primary analytical purpose of a scatter plot?

Scatter plots identify correlations (positive, negative, non-linear) between two continuous measures and locate isolated outlier data points.

How do opacity and borders help with overlapping marks?

Lower opacity creates darker shading where points overlap to show density, while borders keep individual mark boundaries discernible.

Can I encode both Color and Shape using the same dimension?

Yes, dropping a dimension onto Color and Shape simultaneously provides redundant encoding, making categories easier to identify across the chart.

How was this lesson?

Your feedback helps us refine explanations and catch bugs.