How to Validate Data in Tableau

Building a powerful dashboard in Tableau means nothing if the data powering it is flawed. An oversight in data accuracy can lead to misguided business decisions, broken trust with stakeholders, and hours of frustrating troubleshooting. This guide provides practical methods to validate your data directly within Tableau, ensuring you can build and share your analyses with complete confidence.

Why Does Data Validation in Tableau Matter?

Before jumping into the "how," it's helpful to understand the "why." Data validation is the process of ensuring your source data is accurate, clean, and consistent before you start building visualizations. Skipping this step is like building a house on a shaky foundation. You might create stunning charts, but if they're based on incorrect numbers, duplicated records, or mismatched categories, the beautiful facade will quickly crumble under scrutiny.

Think about these common scenarios:

• Reporting a 15% increase in sales, only to find out later that some transactions were counted twice.
• Presenting customer demographics where "California," "CA," and "california" are treated as three separate states.
• Making budget decisions based on an incomplete data set that was missing the last week of the month.

These issues erode credibility and can have serious financial consequences. A few simple validation checks upfront can save you from major headaches down the line.

Start on The Data Source Page

Your first opportunity to inspect your data is a single click away, right after you connect it. The Data Source page in Tableau is your first line of defense against messy data, and spending a few minutes here can reveal a lot.

1. Check Your Data Types

When Tableau connects to your data, it makes an educated guess about the type of data in each column. Sometimes it gets it wrong. A numeric ID might be classified as a number when it should be a string, or a date field might come in as a string.

Scan the icons above each column header:

• 🌐: Geographic data (Country, State, City)
• 📅: Date data
• 📅 & 🕰️: Datetime data
• Abc: String or text data
• #: Numeric data

If you see a date field labeled "Abc" or a numeric product ID with a # icon, click the icon and change it to the correct type. Incorrect data types can cause calculation errors and prevent you from building proper time-series charts or summaries.

2. Review the Data Grid

The grid on the Data Source page shows you a preview of your first 1,000 rows. Don't underestimate the power of a quick visual scan! Scroll through and ask yourself:

• Are there a lot of null values where you expect to see data?
• Does the data's format look consistent? For example, are all dates formatted as MM/DD/YYYY?
• Do categorical values look clean? Look for typos or casing inconsistencies (e.g., "Shipped" vs. "shipped").

3. Check the Record Count

Tableau shows the number of rows from your data source right above the preview grid. This is a simple but powerful check. If your source file is a spreadsheet and has 10,500 rows, but Tableau is only showing 1,200, something may be wrong with the connection or data structure.

4. Use the Data Interpreter

If you're connecting to an Excel or Google Sheets file that is formatted for humans (with merged cells, titles, or empty rows), Tableau can get tripped up. Tableau's Data Interpreter is designed to clean up these kinds of files automatically. If you connect to a spreadsheet and the data looks jumbled, check the "Use Data Interpreter" box at the top of the left pane.

Using Worksheets for Practical Validation

The Data Source page is great for a first glance, but you'll do the bulk of your validation work within actual worksheets. Creating simple views allows you to aggregate the data and spot problems that aren't visible on the row-level data.

Create Summary Tables for Grand Totals

The quickest way to validate your key metrics (like Sales, Revenue, or Cost) is to compare the Tableau grand total to a trusted number from your source system.

1. Open a new worksheet and name it "Validation Totals."
2. Drag the measure you want to check (e.g., Sales) onto the Text card.
3. This shows you the total sum of sales for all records in your data set.

Compare this single number against the official total from your CRM, eCommerce platform, or accounting software. If they're miles apart, something is wrong with the data you're connected to or filters you've applied.

Drag Dimensions to Spot Mismatches

Misspellings and inconsistencies in categorical data are common culprits. A simple bar chart can quickly show you if a dimension needs cleaning. For example, say you have a region column with "Western," "Midwest," "East Coast," and "East." You can easily spot these problems:

1. Open a new worksheet.
2. Drag the Region dimension onto the Rows shelf.
3. Drag Number of Records onto Columns.

Tableau will create a bar chart showing the count of records for each region. Immediately, you'll see that "East" and "East Coast" should probably be grouped together. This visual check makes it much easier than scrolling through thousands of rows to find these errors manually.

Create a Highlight Table to Find Nulls

Null values (blank cells) are especially tricky in important dimensions like customer segment or product category. A highlight table uses color to draw your attention to these potential issues.

1. Drag the dimension of interest to the Rows shelf (e.g., "Product Category").
2. Drag another dimension to Columns (e.g., "Product Name").
3. Drag "Number of Records" to the Text mark, and also drag it to the Color mark.

By default, Tableau will color the cells with non-null values differently, making them stand out at a glance.

More Advanced Validation with Calculated Fields

Sometimes you need more sophisticated logic to uncover deeper problems. Calculated fields are your solution.

Using ISNULL() to Flag Missing Data

You can create a binary flag to separate records with missing values in a specific column.

Say you want to count all orders missing a priority tag:

1. Create a new calculated field and call it "Missing Priority."
2. In the formula editor, enter:

IF ISNULL([Priority]) THEN "Missing" ELSE "Not Missing" END

1. Drag this new field onto Rows and "Number of Records" onto Columns. You now have a clear count of records with and without a missing value.

Validating Date Ranges

A classic error in transactional data is future dates. If your data set ends December 2022, but you see an order from March 2023, something's wrong.

1. Create a calculated field named "Invalid Dates."
2. In the formula editor, enter:

IF [Order Date] > TODAY() THEN "Future Date" ELSE "Valid Date" END

You can use this calculated field as a filter to exclude any future dates, or as an indicator to see how many invalid records there are.

Building a Data Validation Dashboard

To take this a step further, consider creating an entire dashboard dedicated solely to data validation tasks. Add new worksheets to your validation dashboard, each designed to check different aspects of your data:

• A sheet for grand totals
• A sheet highlighting nulls
• A sheet for validating date ranges
• A chart to check for misspelled categories

This approach makes validation a repeatable process. Every time you refresh your data source, you can quickly glance at your validation dashboard and confirm that everything is still okay.

Create a Single Source of Truth

Validation isn't done in a vacuum, it's crucial to connect and verify with your original data sources. Knowing that your Tableau numbers are correct, you can compare them to the original data source. This is the only way to ensure true accuracy.

Pull a Simple Report from the Source

Go to the original platform (e.g., Google Analytics, Shopify, or Salesforce) and pull a simple report that mirrors the view you created in Tableau. Check:

• Sum of sales from Sales Order data per quarter
• Number of sessions from Google Analytics for last year

Create the Exact Same View in Tableau

In Tableau, build the exact same view, applying the same date filters, category filters, and demographics you used in the source report. This allows you to create an apples-to-apples comparison between the two platforms.

What Happens if the Numbers Don't Match

If the validated totals don't match the numbers in your source report, don't panic. This is often the most valuable part of validation as it directs you to issues that need resolution. Common causes include:

• Date Range Discrepancy: Are both reports looking at the same timeframe, e.g., last quarter versus last 90 days?
• Filter Mismatch: Is your source report filtered for a certain product line but your Tableau report is not?
• Definition Differences: How is revenue defined by each platform? Does one include shipping costs, and the other doesn't?
• Joining Issues: If you're joining multiple tables in Tableau, an incorrect join could be creating duplicates or dropping records entirely.

Final Thoughts

Data validation in Tableau is not just a technical checklist, it's a foundational step in building trust in your analyses. By spending time on the Data Source page, creating simple worksheets, and cross-referencing with your original data source, you ensure that the stories you tell with your data are backed by rock-solid facts. That extra manual step might seem tedious, but it's also essential, especially when managing multiple data sources. Using a platform like Graphed (target="_blank" rel="noopener"), where you have hands-on control of your data pipeline and understand the semantic layer of each source, you can spend less time worrying about inconsistencies and more time simply asking questions in plain English to get real-time accurate dashboards.