Learning AI Tools in Tableau

Chapter 4. Tableau Pulse for End Users

A Note for Early Release Readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 4th chapter of the final book. Please note that the GitHub repo will be made active later on.

If you have comments about how we might improve the content and/or examples in this book, or if you notice missing material within this chapter, please reach out to the editor at [e-mail].

In previous chapters, you learned how to enable Tableau Pulse and build both basic and advanced metric definitions. Now it’s time to see how they’re consumed from your audience’s perspective. In this chapter, you’ll see how metrics you create are displayed to end users, what happens when they follow multiple metrics, and how they can receive summarized digests of the AI insights generated.

Pulse Home Page

Just like you saw in Chapter 2, after navigating to Tableau Pulse from Tableau Cloud, end users are greeted with a summary page that shows the metrics they already follow and those that are available for following. However, instead of being able to see all metrics in Pulse, remember they will only have access to those tied to published data sources that they have view and connect permissions for. Figure 4-1 shows the available metric definitions for the Viewer role with access to the published data source used in previous examples, Superstore Transactions.

Figure 4-1. Available metric definitions for a Viewer

It is from this screen that users will access all metric definitions and existing metrics and begin the process of constructing their own metrics. By clicking the triple-dot Actions menu on the right, they’ll be able to access two subsequent areas:

Insights Exploration

The summary and breakdown screen for the metric definition

See All Metrics

A screen with all existing metrics based on the metric definition

They can also click directly on the metric definition itself to navigate to the screen with all metrics.

Once again, this should all seem very familiar, as the actions available are a subset of those you saw in Chapter 3. However, for a Viewer, they can view (hence the name) but cannot modify or delete any metric definitions they have access to.

Note

Although not shown in the figures, if your Tableau Cloud site has the Data Catalog enabled, data quality warnings applied to data assets will show up in the Browse Metrics section at the bottom of a metric card in the Following section, and beneath the metric’s name when viewing either the Overview or Breakdown.

Figure 4-2 shows the available metrics based on the Sales metric definition that was constructed in Chapter 2. You’ll notice that the Viewer cannot see the number of followers for each metric, and as expected, they aren’t able to edit the metric definition.

Figure 4-2. Metrics available for the Sales metric definition

Warning

While Viewers of metrics aren’t able to delete or modify them, they are able to manage the followers. They can both view the existing followers as well as add and remove them from a given metric.

Creating New Metrics

The metric creation process for an end user is identical to what you saw in Chapter 2. To construct a metric, they can either access the metric definition using the Insights Exploration section or, more likely, they can begin by clicking on a specific metric that they may want to customize further. It is important to note that any new metrics they create need to have at least one follower assigned before leaving the screen. This will ensure that the metric is created and accessible for others.

To start seeing how the AI-generated insight summaries are propagated to end users, you can create your own metrics or follow along using a series of metrics outlined in the sidebar “Additional Metrics for Superstore Transactions” from the Superstore Transactions data source:

• Year to Date Profit Ratio

• Year to Date Average Discount

• Year to Date Number of Customers

• Year to Date Items Sold

• Year to Date Profit

• Year to Date Average Time to Ship

• Year to Date Average Products per Order

• Year to Date Average Profit

• Year to Date Orders Shipped

Additional Metrics for Superstore Transactions

You may want to have additional metrics available to further test Tableau Pulse functionality. The Superstore Transactions data set includes many options for constructing such metrics. Here you will find the metric definitions with suggested filters, number formatting, and more. Those marked with an asterisk are advanced definitions.

Table 4-1. A table of additional metrics using the Superstore Transactions data set

Metric name	Measure and aggregation	Time dimension	Adjustable filters	Number format	Value going up is
Orders Shipped	Distinct Count of Order ID	Order Date	Ship Mode, Region	Number	Favorable
Average Profit	Average of Profit	Order Date	Segment, Region	Currency	Favorable
Average Products per Order*	AVG({FIXED Order ID: COUNTD(Product ID)})	Order Date	Category, Segment	Number	Favorable
Average Time to Ship*	AVG({FIXED Order ID: DATEDIFF(‘day’,Order Date,Ship Date)})	Order Date	Ship Mode, State/Province	Number; day, days	Unfavorable
Profit	Sum of Profit	Order Date	Sub-Category, Product Name	Currency	Favorable
Items Sold	Sum of Quantity	Order Date	Category, Sub-Category	Number; item, items	Favorable
Number of Customers	Distinct Count of Customer ID	Order Date	Region, Segment	Number; customer, customers	Favorable
Average Discount	Average of Discount	Order Date	Segment, State/Province	Percentage	Unfavorable
Profit Ratio*	SUM(Profit)/SUM(Sales)	Order Date	Category, Sub-Category	Percentage	Favorable

Figure 4-3 shows the nine metrics from the sidebar after following them. Followed metrics are organized in chronological following order. Remember, Pulse is heavily dependent on today’s date, so the generated summaries and insights may differ significantly in your environment.

Figure 4-3. Pulse summary after following metrics

Tip

Tableau Cloud sites users with a Tableau+ license have the ability to sort metrics in ascending or descending order by data source name, metric definition, or recently followed. A static goal value can also be applied to metrics with a period-to-date time range. The goal will display underneath the large number with a progress bar and within the line chart as a reference line if the sparkline values are set to show as running total. You can learn more about Tableau+ in Chapter 7.

Depending on what happens when metrics are followed, Pulse will begin to construct a larger insight summary that combines the most interesting insights from the user’s followed metrics. These summaries are constructed every 24 hours and may take at least a day to initially show up for the end user. The following is the AI summary readout from Figure 4-3. Metrics in this summary show up as bold and underlined, with the ability for the user to click on them to navigate directly to the detailed view of the metric:

Annual Profit soared by 37.7% compared to last year. Similarly, Orders Shipped increased by 27.7% over the same period. Items Sold remained stable, with a 26.3% higher performance than last year. Overall, 6 of 9 metrics changed: 3 favorably, 1 unfavorably.

Understanding Generated Insights

As mentioned in Chapter 2, Tableau Pulse relies on the dimension fields specified for adjustable filters to generate insights. Behind the scenes, insight types (refer to Table 2-3) are generated and then scored and ranked to determine which are most statistically significant. These insights appear in both the subsequent Breakdown charts and the automated insight summaries.

End users have the option to provide feedback on the helpfulness of each insight by way of a thumbs up or thumbs down button next to each generated insight. When the insight summary is tagged as helpful users can provide a description of up to 3,000 characters. Similarly, as seen in Figure 4-4, users can tag an insight summary as not helpful, select reasons, and also provide feedback. The options for this type of feedback are as follows:

Biased, toxic, or harmful

Language in the output is offensive in some way.

Inaccurate

The numerical values generated in the output are factually wrong.

Incomplete

The output doesn’t draw a conclusion.

Inappropriate tone or style

Language in the output isn’t professional or is overly colorful.

Other

This is for all other issues with the generated output.

Figure 4-4. Types of feedback that can be selected for insight summaries

This mechanism of feedback doesn’t provide any immediate change to the insights generated. Instead, it is passed to the LLM responsible for generating the summaries (likely by way of prompt instructions).

Warning

There is currently no way to undo marking a generated insight or chart as helpful or not helpful. Since this can have an impact on subsequent behavior, I recommend caution when assigning feedback.

Metric Digests

In addition to the Pulse home screen of followed metrics, users can receive digests of their followed metrics. The default setting for these digests is weekly emails and Slack messages, but the frequency can be adjusted to daily, and users can specify which communication channels they want to receive the summary in. Figure 4-5 shows the options available, accessed by clicking the user icon in the upper right of the Pulse home screen and selecting Preferences. Slack must first be configured within a Tableau Cloud site by an administrator to be available as a channel option. You’ll find out how to configure Tableau Cloud and Slack in Chapter 6.

Figure 4-5. Preferences end users can set for Pulse digests

These digests coincide with updates to the generated AI summaries and cannot be configured to a specific time. Based on observation, it appears that the daily insight summaries are constructed somewhere between 3 a.m. and 4 a.m. Coordinated Universal Time (UTC) and then the digests are subsequently sent out.

Email Digest

Within the email digest, you’ll see the same generated summary of the metrics, stating which metrics have experienced the most interesting changes. Beneath the summary are thumbnails, summaries, and hyperlinks to view the details of each metric. Metrics in this format are displayed in alphabetical order. Figure 4-6 shows a fragment of a daily email digest.

Figure 4-6. A snippet of an email digest

The email also contains footer links for the email recipient to manage the metrics they follow or to adjust their digest frequency settings.

Slack Digest

Users can also enable Slack digests, which allow them to see the same Pulse insights directly in their communication platform. Once a Tableau Cloud administrator enables the integration between Slack and Tableau Cloud via the Tableau for Slack app, each user must add the Tableau app to their Slack environment and authorize the connection between the two applications.

The Tableau app can be added by clicking “Add apps” in the workspace sidebar within the Slack workspace, as shown in Figure 4-7.

Figure 4-7. Slack workspace sidebar with the “Add apps” button highlighted

Once clicked, users will be able to select the Tableau app from the list of available apps and begin the authorization process with Tableau Cloud. Authorization requires users to sign in to Tableau Cloud and allow Slack to perform actions on behalf of the user in Tableau Cloud. Figure 4-8 shows the authorization screen with the explicit list of information and actions Slack will be granted.

Figure 4-8. List of authorizations given to Slack when adding Tableau app

Tip

If the Tableau app isn’t already available to users in your Slack workspace and you don’t have permissions to add it, you can initiate a request to your Slack administrator following the same process of attempting to add the app. You’ll end up in an approval request screen with the option to include more information about your request. Slackbot will direct message (DM) you once your request is reviewed by an administrator.

Once the Tableau app is enabled for a user, they will begin to receive Pulse digests at their desired frequency. These come in the form of DMs from the Tableau app in Slack. These digests mirror what is seen in the email digests, starting with the general summary of all metrics followed. Each followed metric is recreated with its current value, comparison value(s), sparkline, and insights. Similar to the email digest, metrics here are shown in alphabetical order. If a user follows many metrics (more than four), the Tableau app will include remaining metrics in a threaded reply. Figure 4-9 shows an example digest.

Figure 4-9. Tableau Pulse digest in Slack

Metrics within these digests have hyperlinks to directly navigate to the insight exploration screen. There are also management buttons identical to the email digest options at the bottom of the message, allowing the user to manage their metrics or digest preferences.

Tableau Mobile

Users can also access Pulse metrics and summaries using the Tableau Mobile app. The Tableau Mobile app is available for free on both Android and iOS, giving users the ability to access their Tableau content on the go. Figure 4-10 shows what Tableau Pulse looks when viewed from an iPhone.

Figure 4-10. Tableau Pulse summary in Tableau Mobile

Within the Tableau Mobile app, users can scroll vertically through their followed metrics (which are displayed in chronological follow order, like on Tableau Cloud). In this regard, the Mobile version is identical to the digests. The user can also interact with the detailed view of the metric and both ask and access top insights. If configured, the user can also receive notifications reminding them to check their Pulse insights. Users can also set Pulse as their starting page when opening the app.

It is important to note that unlike the browser experience, users are unable to create new metrics on their phone. Additionally, they won’t have access to the adjustable filters tied to each metric other than the time horizon.

Comparing the different methods of accessing the digest, the Mobile app experience offers the richest and easiest to use experience. Users can seamlessly toggle between the summary and detailed views of their followed metrics, the metrics themselves are larger and easier to read, and all of the visualizations are interactive. For busy executives, it’s the easiest way to access all their metrics in one place.

Summary

This chapter was an in-depth look at what end users experience when using Tableau Pulse. Here are some key takeaways for what you’ve learned:

• Users with Viewer licenses can construct metrics only based on premade metric definitions.

• All users, regardless of license type, can manage and modify followers of metrics.

• Users can tag insights as helpful or not helpful and provide feedback.

• Users can receive digests via email or Slack on a configurable schedule (daily, weekly, or monthly).

• Both email and Slack digests include hyperlinks to access and manage followed metrics.

• A user must enable the Tableau app for Slack and authorize Slack to begin receiving digests.

• The Tableau Mobile app includes Tableau Pulse and provides the easiest and most immersive experience for accessing followed metrics.

• Users cannot create new metrics or access adjustable filters in the Mobile app.

In Chapter 5, you will see how a variety of industries can benefit from using Tableau Pulse. The chapter will discuss several metric examples and expose you to a variety of data sources. You’ll also see metrics that can be created from the Admin Insights data sources mentioned in Chapter 3, giving you tools to monitor usage and adoption of Pulse across your Tableau site.