Configure Tableau Data Story Settings: Analytics

Important changes for Tableau Data Stories

Tableau Data Stories will be retired in Tableau Desktop, Tableau Cloud, and Tableau Server in January of 2025 (2025.1). With advances in natural language technologies, we're developing an improved interface that will make it easier to ask questions of your data and stay on top of changes. For more information, see How Tableau Pulse powered by Tableau AI is Reimagining the Data Experience(Link opens in a new window).

Within your Tableau Data Story, you can choose which analytics to write about and when those analytics are written about. Different types of analytics are available depending on your story type and how many dimensions and measures your story has. However, analytics aren't currently supported for scatter plot story types. For more information, see Choose the Right Story Type for Your Tableau Data Story.

Configure analytics for your story

  1. Add a Tableau Data Story to a Dashboard.
  2. From your dashboard, click the Settings icon at the top-left corner of your Data Story object.
  3. In the Data Story dialog box, click the Analytics tab.
  4. Click the switches to turn on different types of analytics.
  5. For Segments and Trend line, expand Settings to set thresholds for performing those analytics.
  6. Click Save.

Understand different types of analytics

Correlation

Use Correlation to identify true statistical correlations between two series. If you have more than two series, then all combinations are analyzed for correlations. For example, you might turn on Correlation to identify when two products are often purchased together.

Clustering

Use Clustering to identify distinct groups of data points (clusters) using a single statistical analysis. For example, you might turn on Clustering to identify when a product is very popular in a specific geographic region.

Distribution

Use Distribution to rank data points relative to each other using nonstatistical observations, such as mean, median, skew, etc. For example, you might turn on Distribution to identify which product has the highest profit ratio.

Segments

Use Segments to highlight noteworthy changes to data points within a series. First, set the minimum percentage of change that you want to be written about in a segment. Changes that fall below your defined threshold aren't written about. For example, if you set your segment threshold for changes that are greater than 60%, then your story doesn't write about a trough in a time series that features a 30% decrease.

After you’ve set your threshold, choose whether to Apply formatting, and set the minimum percentage of change that you want formatted.

Trend line

Use Trend line to calculate a linear best fit line and identify data that falls within a defined percentage of confidence. Data that has high variability has a lower confidence level than data that's more consistent, and that confidence level affects whether trend lines are written about. You can use trend lines for stories that have one dimension and one measure, or you can use trend lines in a drilldown. For more information about drilldowns, see Configure Tableau Data Story Settings: Narrative.

Set the minimum percentage of confidence for your trend line. If you set your threshold at 95%, but a trend line could be drawn at 90% confidence, then your story doesn't write about trend lines. After you’ve set your threshold, choose whether to Apply formatting. Then set the minimum percentage of change that you want formatted.

Tableau Data Stories about trend lines communicate the absolute change over a period. The story written about your trend line varies depending on the level of verbosity you set for your story. If your story uses high verbosity, then your story writes about the R-squared value, which is a statistical concept that quantifies how well your data fits the trend line. For more information about verbosity settings, see Configure Tableau Data Story Settings: Narrative.

Within the Trend line settings, you can also choose how many periods into the future for which you want your story to write predictions. When you use predictions, your story uses the slope and intercept of the trend line to calculate predicted values for future periods. The confidence of the prediction adds upper and lower bounds to the confidence threshold you set for trend lines. You can use predictions when your story has at least 30 data points that are linear.

Volatility

Use Volatility to analyze standard deviations over time. For example, use Volatility when you want your story to write about values that fall outside the average range for your data.

Break down how analytics are used to generate stories

At this point, you might be wondering how the analytics for different story types work. Let’s take a look at an example for each story type and break down each sentence in the story.

Understand analytics for discrete stories

Because continuous stories measure trends over time, Data Stories writes about performance, progression, averages, totals, streaks, volatility, segments, and predictions.

The following example of a continuous story is about sales per month:

An example story with text insights about sales per month. The text in this image is transcribed in the following table.

Example story Story breakdown
  • Average Sales was $47,858 across all 48 months.
  • The minimum value was $4,520 (February 2014) and the maximum was $118,448 (November 2017).
The first two sentences use average and range functions to write about the average, maximum, and minimum values across the period you’re analyzing.
  • Sales increased by 489% over the course of the series but ended with a downward trend, decreasing in the final month.
The third sentence is about overall performance of the measure over the period. For example, a sentence can be about whether sales increased, decreased, or trended differently during a specific period.
  • The largest single increase on a percentage basis occurred in March 2014 (+1,132%). However, the largest single increase on an absolute basis occurred in September 2014 (+$53,868).
The fourth sentence uses progression analysis. This sentence writes about the largest increase and decrease based on the measure during the period using both a percentage basis and absolute basis.
  • Of the three series, the strongest relationship was between Corporate and Home Office, which had a moderate positive correlation, suggesting that as one (Corporate) increases, the other (Home Office) generally does too, or vice versa.
This sentence is a Correlation insight. This type of analytic insight writes about notable correlations between different series in your data.
  • Sales experienced cyclicality, repeating each cycle about every 12 months. There was also a pattern of smaller cycles that repeated about every three months.
  • Sales had a significant positive peak between October 2014 ($31,453) and February 2015 ($11,951), rising to $78,629 in November 2014.
This sentence is a Segment insight. This type of analytic insight writes about noteworthy increases and decreases over time.
  • The overall linear trend of the series rose at $902 per month for an absolute change of $42,394 over the course of the series. If this trend continued for the next one month, Sales is predicted to be about $69,958.
This sentence is a Trend line insight. This type of insight writes about how well trends fit your data with a certain percentage of confidence, and trend lines allow you to make predictions based on historic trends.

Understand analytics for discrete stories

Because discrete stories allow you to compare values and understand the distribution of the data, the story writes about distribution, averages, totals, and groupings or clusters across the data.

The following example of a discrete story is about sales by product:

An example story with text insights about sales by product. The text in this image is transcribed in the following table.

Example story Story breakdown
  • Total Sales is $2.3 million across all 17 products.
The first sentence calculates the total value of your measure.
  • The Sales of $2.3 million was driven by Phones with $330,007, Chairs with $328,449, and Storage with $223,844.
The second sentence writes about the dimension drivers. In this example, the dimension drivers are the products that contributed the most to total sales.
  • The distribution is positively skewed as the average of $135,129 is greater than the median of $114,880.
  • Sales is relatively concentrated with 78% of the total represented by eight of the 17 products (47%).
The third and fourth sentences analyze the distribution of the data. This analyzes the averages, medians, concentration of data (if any exist), and how the data is skewed. This helps identify how balanced these grouped variables are compared to one another.
  • The top two products combine for over a quarter (29%) of overall Sales.
This sentence uses Clustering to write about measures that can be grouped. This helps identify whether there are distinct groups that stand out in the data.
  • Phones ($330,007) is more than two times bigger than the average across the 17 products.
The final sentence writes about notable outliers.

Understand analytics for scatter plot stories

Scatter plot story types are best used to understand the relationship between two measures, and for that reason, scatter plot stories require 2–3 measures. The scatter plot analysis writes about the relationship (regression) between two measures, and it writes about groups (clusters) within the data, if they exist.

The following example of a scatter plot story is about profit and sales across a dimension:

An example story with text insights about profit and sales. The text in this image is transcribed in the following table.

Example story Story breakdown
  • As quantity increased and profit increased, sales increased based on the data provided. Specifically, when quantity increased by 1, sales increased $49.55, and when profit increased by $1.00, sales increased $1.20.
  • Few customers deviated from this general relationship, indicating a good fit.
The first two sentences are powered by regression analytics. Regression shows how one measure affects another. Notice that in the first sentence, the story has identified a relationship between profit and sales.
  • When organized into groups of similar profit, quantity and sales values, one distinct group stands out. There were 651 customers that had values of profit between -$6,626 and $1,488, quantity between 2 and 122 and sales between $4.83 and $5,690.
The third sentence is derived from clustering. Clustering analytics tries to identify key groups or clusters across all the variables in the data.
  • Tamara Chand, Raymond Buch, and Sanjit Chand, among others were outliers with high profit and sales values. Sean Miller stood out with a low profit and high sales value.
The fourth sentence is written about outliers–values that fall significantly above or below the average.
  • The minimum value for profit is -$6,626 (Cindy Stewart) and the maximum value is $8,981 (Tamara Chand), a difference of $15,608. The average profit per customer is $361 and the median is $228.
  • The minimum value for quantity is 2 (Anthony O'Donnell) and the maximum value is 150 (Jonathan Doherty), a difference of 148. The average quantity per customer is 47.76 and the median is 44.
  • The distribution of sales ranges from $4.83 (Thais Sissman) to $25,043 (Sean Miller), a difference of $25,038. The average sales per customer is $2,897 and the median is $2,256.
The remaining sentences for scatter plot stories use range and average analysis to write insights.

Understand analytics for percent of whole stories

Percent of whole story types are best for understanding what part of a whole a dimension or measure represents.

The following example of a percent of whole story is about sales by segment:

An example story with text insights about sales by segment. The text in this image is transcribed in the following table.

Example story Story breakdown
  • Total SUM(Sales) is 2.3 million across all three entities.

The first sentence calculates the total value of your measure.
  • The SUM(Sales) of 2.3 million was driven by Consumer with 1.2 million, Corporate with 706,146 and Home Office with 429,653.

The second sentence writes about drivers. In this example, the drivers are segments that contributed the most to total sales.
  • The minimum value is 429,653 (Home Office) and the maximum is 1.2 million (Consumer), a difference of 731,748, averaging 765,734.
The final sentence analyzes the distribution of the data.
Thanks for your feedback!Your feedback has been successfully submitted. Thank you!