How to Visualize Correlation in Tableau
Seeing how two variables relate to each other is fundamental to making smart business decisions. Whether you're asking, "Does more ad spend actually lead to more sales?" or "Are higher customer ratings linked to fewer support tickets?", the answer lies in understanding correlation. This tutorial will walk you through how to create two of the most powerful charts for visualizing these relationships in Tableau: the scatter plot and the heatmap. We'll cover each method step-by-step so you can start turning a flat table of numbers into actionable insights.
Why Visualizing Correlation Matters
Correlation shows you the relationship between two or more variables. It tells you if they move in the same direction (positive correlation), in opposite directions (negative correlation), or have no clear connection at all. For example, the hours you spend studying for a test likely have a positive correlation with your final grade, as one goes up, the other tends to follow.
While a statistician can give you a correlation coefficient - a single number like r = 0.8 - a visualization does something more powerful. It tells a story. A chart can reveal outliers, show clear clusters or groups in your data, and expose non-linear relationships that a single number would completely miss. Turning data into a picture makes complex relationships intuitive and easy for anyone to understand, not just data experts.
Method 1: Building a Scatter Plot in Tableau
The scatter plot is the go-to chart for visualizing the relationship between two numerical variables. Each dot (or 'mark') on the chart represents a single data point, plotted along a vertical and horizontal axis. It's the cleanest way to see if a pattern emerges from the chaos.
Step-by-Step Guide to Creating a Scatter Plot
Let's imagine we're analyzing data from a superstore, and we want to see the relationship between the Sales of an order and its Profit. Does a higher sale price always mean higher profit? Let's find out.
Step 1: Add Your Measures to the View
Measures are your numerical data fields in Tableau. In this case, our measures are Sales and Profit.
Drag the Sales measure from your data pane and drop it onto the Columns shelf.
Drag the Profit measure and drop it onto the Rows shelf.
What you’ll see is a single dot on the screen. This is happening because Tableau is helpful by default, automatically aggregating your measures. Right now, it’s showing you the sum of ALL sales versus the sum of ALL profit. We want to see how each individual transaction looks, not the grand total.
Step 2: Disaggregate Your Data to See All Points
To break that single point into a "scatter" of individual points, you need to tell Tableau to stop aggregating the measures.
Go to the Analysis menu at the top of the Tableau window.
Uncheck the option that says Aggregate Measures.
Suddenly, the canvas will fill with points. Each point now represents a single row in your dataset (for instance, an individual order). You’ve just created your first scatter plot!
Step 3: Add Context with Dimensions
A basic scatter plot is good, but a great scatter plot tells a richer story. You can use your categorical data fields (Dimensions) to add layers of meaning. For example, is the sales-profit relationship different for certain product categories?
Find the Category dimension on your Data pane.
Drag Category and drop it directly onto the Color part of the Marks card.
Now, each dot is colored based on the product category it belongs to. Instantly, you can see if certain categories (like Technology, in blue) tend to have higher sales and profits compared to others (like Office Supplies, in orange).
Interpreting Your Scatter Plot
Now that your chart is built, it's time for the analysis. Look at the overall trend of the dots:
Positive Correlation: The points form a pattern that trends upward from left to right. This means that as one variable increases, the other tends to increase as well. In our example, we can see a generally positive correlation between sales and profit.
Negative Correlation: The dots slope downward from left to right. As one variable increases, the other tends to decrease. For example, you might see this if you plotted discount amount vs. profit margin.
No Correlation: The points are spread out randomly with no discernible pattern. The variables don't seem to have any relationship.
You can also spot outliers. In our chart, you might notice some points with very high sales but negative profit. These are worth investigating! What specific orders or products are losing money despite selling for a high price?
Adding a Trend Line for Clarity
To make the overall relationship crystal clear, you can add a trend line. It summarizes the primary direction of your data points.
Navigate to the left-hand sidebar and click on the Analytics pane (next to the Data pane).
Drag Trend Line from the pane and hover over your chart.
Drop it on the Linear model option.
Tableau will draw a line that best fits your data. When you hover over it, you'll see a regression formula, an R-squared value, and a p-value. In simple terms, R-squared tells you how much of the change in profit can be explained by the change in sales. A low p-value (typically under 0.05) tells you the relationship is statistically significant and likely not due to random chance.
Method 2: Using a Heatmap to Show Correlation
While a scatter plot is perfect for comparing two measures, what happens when you want to see how one measure performs across the intersection of two dimensions? A heatmap (often called a highlight table in Tableau) is an excellent choice. It uses color intensity to represent values, making it incredibly easy to spot "hot spots" and "cold spots" in your data at a glance.
Let's say we want to find out which product sub-categories have the highest sales in different regions.
Step-by-Step Guide to Creating a Heatmap
Step 1: Set Up the Rows and Columns
Start by placing your dimensions on the rows and columns shelves to create the structure of your table.
Drag the Sub-Category dimension to the Rows shelf.
Drag the Region dimension to the Columns shelf.
You'll see a table filled with "Abc" placeholders. This is the basic grid for your heatmap.
Step 2: Add Color to Represent Value
Now, let's use color to represent our measure: Sales.
Go to the Marks card. It's probably set to Automatic. Click the dropdown menu and select Square. This tells Tableau you want each cell in your grid to be a colored box.
Next, drag the Sales measure onto the Color part of the Marks card.
Tableau will color each square based on its sales volume. By default, it uses a blue gradient where darker shades represent higher sales. You can now instantly see where the most promising combinations are.
Step 3: Add Labels for Precision
The colors give you a great overview, but sometimes you need to see the exact numbers. It’s easy to add them.
Drag the Sales measure again, this time dropping it onto the Label part of the Marks card.
The sales figures will now appear inside each square. This gives you the best of both worlds: a quick visual scan using color and precise numerical data for deeper analysis.
Interpreting Your Heatmap
Reading a heatmap is all about spotting patterns in the colors. In our example, you can quickly identify that the "Phones" and "Chairs" sub-categories have consistently strong sales across all regions, indicated by their dark blue squares. Conversely, categories like "Fasteners" and "Labels" are much lighter, showing lower sales volumes. You're not looking at a statistical correlation here, but you are visually correlating a metric's performance across dimensions, allowing you to instantly identify your strongest and weakest segments.
Putting It Into Practice: Which Chart Should You Use?
Both charts are fantastic for analyzing relationships, but they serve different purposes. Here's a quick guide on when to use each:
Use a Scatter Plot when:
You are comparing two numeric variables (e.g., Marketing Budget vs. Revenue).
You need to see the strength and direction (positive, negative) of the relationship.
You want to identify individual outliers that deviate from the general pattern.
Use a Heatmap (Highlight Table) when:
You are comparing a numeric variable across two categorical variables (e.g., Sales per Product per City).
You need a quick, high-level overview to spot the best and worst performing combinations.
You're presenting to a broad audience who needs an immediate, intuitive visual.
Final Thoughts
Visualizing correlation in Tableau transforms raw data into a clear story. By mastering scatter plots and heatmaps, you're no longer just looking at spreadsheets - you're uncovering the hidden patterns that drive your business. Whether you're tracking two critical metrics or scanning for performance hotspots, these visualizations are essential tools for anyone looking to make more data-informed decisions.
Learning the ins and outs of powerful BI tools like Tableau is a fantastic skill, but it often requires a big-time investment in learning and setup. At Graphed, we're focused on removing that complexity by letting you ask questions about your data in plain English. For those times when you need to quickly understand your cross-channel marketing performance or get sales numbers in seconds, we help provide answers without the hassle of navigating complex interfaces. You can learn more and try it yourself at Graphed.