Get Started with Explain Data
Use Explain Data in your flow of analysis as you are exploring the marks in a viz. The best way to get started with Explain Data is to select a mark, run Explain Data and start exploring explanations.
Use Explain Data
Author Workbooks and Control Access
The basic steps to run Explain Data are:
- Select a mark in a viz.
- Hover the cursor over the mark, and then click Explain Data in the tooltip menu.
The Explain Data pane opens with possible explanations for the value of the explained mark. Click different explanation names to expand the details and start exploring.
Tips for using Explain Data
- You must select a single mark – only one mark can be explained at a time.
- The view must contain marks that are aggregated using SUM, AVG, COUNT, COUNTD or AGG (a calculated field).
- If Explain Data cannot analyse the type of mark selected, the Explain Data icon and context menu command will not be available. For more information, see Situations where Explain Data is not available.
- The data you analyse must be drawn from a single, primary data source. Explain Data does not work with blended or cube data sources.
- For information on what characteristics make a data source more interesting for use with Explain Data, see Requirements and Considerations for Using Explain Data.
In the Explain Data pane, click on an explanation name to see more details.
Click on the arrows to expand or contract explanations.
Scroll to see explanation details.
Hover over charts in the explanations to see more detail. Click the Open icon to see a larger version of the visualisation.
Creators or Explorers who open the view for editing can click on the Open icon to open the visualisation as a new worksheet and explore the data further.
Note: Creators and Explorers who have editing permissions can also control Explain Data Settings. For more information, see Control Access to Explain Data.
- Hover over a Help icon to see tooltip help for an explanation. Click on the Help icon to keep the tooltip open. Click on a Learn More link to open the related help topic.
When you run Explain Data, the explanations that are presented in the Explain Data pane specifically apply to the mark you selected. If you click on a different mark, unselect the explained mark or navigate to a different sheet in the workbook, a Reselect button appears in the view thumbnail image at the top of the Explain Data pane.
If you click Reselect, Tableau returns you to the original view and worksheet and reselects the explained mark. Click Update to run Explain Data again.
To explore a new mark, click on another mark, and then click Run in the Explain Data pane.
Explain Data might also display messages to indicate the view has changed (such new fields or filters added or removed from the view), if the data source has changed or if Explain Data settings have changed.
- Run Explain Data on a mark.
- In the Explain Data pane, under Contributing to the value of, click on a measure name.
- Click on the number-of-fields link at the bottom of the pane.
Authors have the option to open Explain Data Settings to control which fields are included in the analysis. For more information, see Change fields used for statistical analysis(Link opens in a new window).
The following terms and concepts appear frequently in explanations. You may find it helpful to become acquainted with their meaning in the context of using Explain Data.
A mark is a selectable data point that summarises some underlying record values in your data. A mark can be made of a single record or multiple records aggregated together. Marks in Tableau can be displayed in many different ways such as lines, shapes, bars and cell text.
Tableau gets the records that make up the mark based on the intersection of the fields in the view.
The explained mark refers to a mark that you selected in the view that was analysed by Explain Data.
For more information on marks, see Marks(Link opens in a new window).
The expected value for a mark is the median value in the expected range of values in the underlying data in your viz. The expected range is the range of values between the 15th and 85th percentile that the statistical model predicts for the explained mark. Tableau determines the expected range each time it runs a statistical analysis on a selected mark.
If an expected value summary says the mark is lower than expected or higher than expected, it means the aggregated mark value is outside the range of values that a statistical model is predicting for the mark. If an expected value summary says the mark is slightly lower or slightly higher than expected or within the range of natural variation, it means the aggregated mark value is within the range of predicted mark values, but is lower or higher than the median.
For more information, see What is an expected range?(Link opens in a new window)
Each column name in a database is a field. For example, Product Name and Sales are both fields. In Tableau, fields like Product Name that categorise data are called dimensions; fields with quantifiable data like Sales are called measures. Tableau aggregates measures by default when you drag them into a view.
Some explanations describe how the underlying record values and the aggregations of those values may be contributing to the value of the explained mark. Other explanations may mention the distribution of values across a dimension for the explained mark.
When you run Explain Data on mark, the analysis considers dimensions and measures in the data source that aren't represented in the view. These fields are referred to as unvisualised dimensions and unvisualised measures.
For more information on dimensions and measures, see Dimensions and Measures(Link opens in a new window).
An aggregate is a value that is a summary or total. Tableau automatically applies aggregations such as SUM or AVG whenever you drag a measure onto Rows, Columns, a Marks card option or the view. For example, measures are displayed as SUM(Sales) or AVG(Sales) to indicate how the measure is being aggregated.
To use Explain Data, your visualisation must use a measure that is aggregated with SUM, AVG, COUNT, COUNTD or AGG.
For more information about aggregation, see Data Aggregation in Tableau(Link opens in a new window).
A record is a row in a database table. A row contains values that correspond to each field. In this example, Category, Product Name and Sales are fields (or columns). Furniture, Floor Lamp and $96 are the values.
A distribution is a list of all the possible values (or intervals) of the data. It also indicates how often each value occurs (frequency of occurrence).