How to Remove Outliers in Tableau

Spotting a data point that throws off your entire Tableau dashboard can be frustrating. These outliers can warp your averages, flatten your trendlines, and lead you to draw the wrong conclusions from an otherwise perfectly good visualization. Getting a clearer picture often means knowing how to identify and handle them. This article will show you exactly when you should (and shouldn't) remove outliers and walk you through three practical methods for filtering them out in Tableau.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

What Are Outliers and Why Do They Matter?

In simple terms, an outlier is a data point that is significantly different from other observations in a dataset. Imagine you’re analyzing daily sales data for a coffee shop. Most days, you sell between $800 and $1,200. But one day, due to a massive catering order for a local conference, you record a sale of $10,000. That $10,000 day is an outlier.

While that massive sale is great for business, it can cause problems in your analysis for a few reasons:

• Distorted Averages: If you calculate your average daily sales for the month, that single $10,000 day will dramatically pull the average up, giving you a misleading idea of what a "typical" day looks like.
• Skewed Visualizations: If you plot daily sales on a line chart, the Y-axis will adjust to accommodate the $10,000 peak. This makes all your other daily sales - the normal $800 to $1,200 range - look like a nearly flat, uninteresting line at the bottom of the chart. It completely hides the actual day-to-day fluctuations.
• Flawed Predictions: If you use this data to forecast future sales, the outlier might lead a statistical model to predict unrealistically high sales for future periods.

Effectively managing outliers helps you build more accurate, readable, and insightful dashboards. It ensures your story about the data is focused on the real patterns, not the rare exceptions.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

To Remove or Not to Remove? That Is the Question

Before you start filtering data, it's important to pause and ask why the outlier exists. Not all outliers should be discarded. Your decision to keep or remove them depends on their context.

When You Should Consider Removing Outliers

• Data Entry Errors: The most common reason for bad outliers. Someone accidentally typed a customer’s age as 150 instead of 50, or added an extra zero to a sales transaction ($5,000 instead of $500). These are mistakes and should be corrected or removed.
• Measurement Errors: A faulty sensor, a glitch in a tracking script, or a broken piece of equipment can produce erroneous data. If you're tracking website load times and a server issue caused a 3-minute load time, that’s not a reflection of typical user experience and can be removed.
• To Understand the "Normal" Case: If your goal is to understand the typical customer journey or the standard performance of a sales rep, you might temporarily exclude an extremely unusual one-off event. It's not that the event isn't real, but it's not representative of the pattern you're trying to analyze.

When You Should Keep (and Investigate) Outliers

• They Represent Real, Valid Events: That $10,000 catering order from the coffee shop example? That's real revenue. Removing it in a report for an executive would be hiding a significant win. The same goes for a genuine, high-value transaction from a whale client or a massive traffic spike from a viral social media post. These aren't errors, they are critical business events worth studying.
• The Outlier Is the Insight: Sometimes, the most interesting stories are in the exceptions. A spike in negative product reviews could signal a manufacturing defect. An unusually high number of support tickets might point to a site outage. A salesperson with sales numbers far exceeding everyone else might be a superstar whose techniques should be taught to the rest of the team. Ignoring these outliers means missing major risks or opportunities.

Method 1: The Simple Drag-and-Drop Filter

The fastest way to remove an outlier in Tableau is by excluding it directly from a visualization. This method is best for one-off analyses or when you have a clear, obvious outlier you want to remove quickly.

Let's say you have a scatter plot showing Sales vs. Profit for all your Customers. You notice one customer with extremely high sales that is stretching the axis of your chart.

1. Identify the Outlier: Locate the data point on your chart that represents the outlier. You can hover over it to confirm the details in the tooltip.
2. Select and Exclude: Simply click on that single data point to select it. A tooltip will appear. In that tooltip, you will see an option to "Exclude." Click it.
3. The Change: Tableau instantly creates a filter for the Customer Name dimension, removing the one you selected. Your viz will rescale, giving you a much clearer view of the remaining data points.

Pros: Extremely fast, intuitive, and requires no code at all. Cons: It’s a manual process. If your data updates and a new outlier appears, you'll have to repeat the process. It's also subjective, as you are "eyeballing" what counts as an outlier.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

Method 2: Using a Calculated Field to Define Outliers

For a more dynamic and statistically sound approach, you can create a calculated field that automatically identifies which data points are outliers. This is the most robust and replicable method. Here are two popular statistical approaches.

The Standard Deviation Method

A common rule of thumb is that any data point that falls more than 2 or 3 standard deviations away from the average is considered an outlier. We can translate this logic into a Tableau calculation.

Here’s the breakdown:

1. Create a Calculated Field: Go to the drop-down menu in the top-left Data pane and select "Create Calculated Field." Name it something like "Outlier Flag (StDev)."
2. Enter the Formula: This formula checks if a customer's total sales fall outside a range of 2 standard deviations from the average sales of all customers in the view.
3. Apply the Filter: Drag your new "Outlier Flag (StDev)" calculated field onto the Filters shelf.
4. Filter Out the Outliers: A dialog box will appear. Select "False" and click OK. This tells Tableau to only keep the data points that are not outliers. Your view will update automatically.

The Interquartile Range (IQR) Method

The IQR method is often considered more robust because it's less affected by the extreme values of the outliers themselves. It defines an outlier based on its distance from the central 50% of the data.

An outlier is any point that falls below Q1 - 1.5IQR or above Q3 + 1.5IQR, where:

• Q1 is the first quartile (the 25th percentile).
• Q3 is the third quartile (the 75th percentile).
• IQR is the Interquartile Range (Q3 - Q1).

You can create this in Tableau with a few calculated fields:

1. Create Q1 Calculation: Name it "Q1" and use the formula:
2. Create Q3 Calculation: Name it "Q3" and use the formula:
3. Create IQR Calculation: Name it "IQR" and use this simple formula:
4. Create Outlier Flag: Finally, create the main flag. Name it "Outlier Flag (IQR)" and use this boolean logic:
5. Apply the Filter: Just like before, drag your "Outlier Flag (IQR)" field to the Filters shelf and select "False" to exclude the outliers.

This method is the logic that Tableau uses to create the "whiskers" on a box plot, so it’s an excellent and standard way to approach outlier detection.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Method 3: Using Sets and Parameters for Adjustable Filtering

What if you want to give yourself (or other users of your dashboard) the ability to define what an outlier is on the fly? You can accomplish this with a combination of a Set and a Parameter.

This approach lets you create an interactive control to adjust the outlier threshold.

1. Create a Parameter: In the Data pane, right-click and select "Create Parameter." Let’s call it "Sales Threshold." Set the Data type to Float, and set a current value like 20,000. This parameter will act as our adjustable ceiling for what counts as a normal sale.
2. Create a Set: Find the dimension you are analyzing (e.g., "Customer Name"), right-click it, and select Create > Set. Name it "Outlier Customers."
3. Define the Set's Condition: In the Create Set window, go to the "Condition" tab. Select "By formula" and enter a condition that uses your new parameter:
4. Apply as a Filter: Drag your new "Outlier Customers" set to the Filters shelf. By default, Tableau will only show the members in the set. To show everything else, right-click the filter and select "Show In/Out," then choose "Out."
5. Show the Parameter: Right-click your "Sales Threshold" parameter in the Data pane and select "Show Parameter." An interactive box will appear on your dashboard, allowing you to instantly type in a new number and watch the visualization update.

This method gives you complete control, blending the power of a dynamic rule with the flexibility of manual adjustment.

Final Thoughts

Dealing with outliers is a common task in data analysis, and knowing how to handle them is a valuable skill for anyone building reports. Remember to first understand the context behind an outlier before deciding to filter it out. With methods ranging from a simple click to dynamic calculated fields, Tableau provides the right tools for a cleaner, more accurate visualization.

While mastering these techniques is powerful, it often takes time to set up and remember the formulas. For when you just need a clean chart without spending 20 minutes writing calculations or configuring sets, we wanted to build a faster way. With Graphed, you simply connect your data sources like Google Analytics or your CRM and describe what you need in plain English. You could ask to “show me ad spend vs. revenue but filter out the top 1% to remove outliers” and our AI handles building the live dashboard for you in seconds. You get to the insight without getting lost in the setup.