Applying Advanced Analytics

The previous two chapters focused on foundational knowledge regarding Power BI report planning, design, and how to create and format visualizations. This chapter builds upon that foundation to demonstrate how the advanced analytical capabilities of Power BI can create an even more compelling user experience and assist users in exploring data to derive insights. This includes a review of artificial intelligence (AI) powered visuals, custom visuals, animation, and other analytics features with Power BI Desktop and the service.

In this chapter, we will review the following topics:

• AI visuals
• ArcGIS Maps for Power BI
• R and Python visuals
• Custom visuals
• Animation and data storytelling
• Analytics pane
• Quick insights/Analyze
• Mobile-optimized report pages

AI visuals

Artificial intelligence (AI) visuals incorporate aspects of machine learning such as pattern recognition and natural (human) language parsing. Machine learning is a form of AI that uses statistical models and algorithms to analyze and make inferences from complex data patterns.

In the past, the application of AI or machine learning required a data scientist, but Microsoft has made tremendous strides in democratizing machine learning within Power BI both with the Auto ML features of dataflows in the Power BI service and AI visuals within Power BI Desktop. The use of the advanced capabilities of AI visuals can add analytical punch to many reports.

Power BI Desktop contains four AI visuals as follows:

• Key influencers
• Decomposition tree
• Q&A
• Smart narrative

We explore each of these visuals in turn, starting with the key influencers visual.

Key influencers

The key influencers visual is a powerful analytic tool included...

R and Python visuals

The R and Python programming languages both have strong data analysis and visualization capabilities. This makes these languages a natural fit for data analysis and visualization tools such as Power BI Desktop. In addition to other integration options, such as the ability to use R and Python in Power Query, Power BI Desktop also provides standard visuals for both R and Python.

In order to use the R and Python visuals, you must first install the R and Python programming languages and ensure that the R scripting and Python scripting options are configured in the GLOBAL options (File | Options and settings | Options) as shown in Figure 8.9:

Figure 8.9: R scripting options

As shown in Figure 8.9, Power BI attempts to detect any installed R and Python program language installations and integrated development environments (IDEs). Once you have the R and Python languages installed and configured, you are ready to add an R or Python visual to a report...

ArcGIS Maps for Power BI

The ArcGIS Map visual for Power BI enables report authors to develop map visualizations far beyond the capabilities of the bubble and filled map visuals described in Chapter 7, Creating and Formatting Visualizations. Created by Esri, a market leader in Geographic Information Systems (GIS), the ArcGIS Map supports all standard map types (for example, bubble and heatmap), but also provides many additional features including a clustering map theme for grouping individual geographic points and the ability to filter a map by the points within a geographical area.

The ArcGIS Map also enables deep control over the logic of the size and color formatting, such as the number of distinct sizes (classes) to display and the algorithm used to associate locations to these classes. Additionally, reference layers and cards of demographic and economic information are available to provide greater context and advanced analytical insights.

The ArcGIS Map visual is included...

Custom visuals

In addition to the standard visuals included in the Visualizations pane of Power BI Desktop, a vast array of custom visuals can be added to reports to deliver extended functionality or to address specific use cases such as applying ‘smooth lines’ formatting to a line chart or displaying multiple KPIs in the same visual via the Power KPI visual by Microsoft.

These visuals, many of which have been created by Microsoft, are developed with the common framework used by the standard visuals and are approved by Microsoft prior to inclusion in Microsoft AppSource. Given the common framework, custom visuals can be integrated into Power BI reports with standard visuals and exhibit the same standard behaviors such as filtering via slicers and report and page filters.

Power BI report authors and BI teams are well-served to remain conscious of both the advantages and limitations of custom visuals. For example, when several measures or dimension columns need...

Animation and data storytelling

A top responsibility for many data professionals is the ability to convey their findings to others in a clear and compelling fashion. Common scenarios for data storytelling include recurring performance review meetings (for example, the close of a fiscal period) and special project or ad hoc meetings with senior managers and executives. For these meetings, the data professional or team has already identified the analytical insights to highlight but must plan to properly communicate this message to the specific stakeholders or audience.

Power BI animation features, including bookmarks as described in Chapter 6, Planning Power BI Reports, provide powerful support for data storytelling. In addition, the play axis available in the standard Scatter chart visual and the animation features available in many custom visuals, such as the Line Dot chart and the Pulse chart, can also be used to deliver advanced analytical insights and data storytelling.

...

Analytics pane

In addition to the Visualization pane’s Build visual and Format visual sub-panes used to create and format report visuals discussed in the previous chapter, an Analytics pane is also available for certain Cartesian standard visuals such as Line charts and Clustered column charts.

The Analytics pane allows report authors to add constant and dynamic reference lines such as average, max, and min to visuals to provide greater context and analytical value. Additionally, trend and forecast lines can be added to display the results of advanced analytical techniques such as exponential smoothing to support predictive analytics.

A simple but important use case of the Analytics pane, exemplified in the Trend line section below, is to add a constant line that represents a goal or threshold to compare a measure against. Dynamic reference lines representing an aggregation (for example, a median) behave just like DAX measures and thus, in some scenarios, avoid the...

Quick insights/Analyze

Quick insights, also known as the Analyze feature, is one of the most analytically advanced features in Power BI as it enables sophisticated machine learning algorithms to be executed against datasets or specific subsets of those datasets. The results of these computations automatically generate highly formatted Power BI visuals that can be integrated into reports as though they were created from scratch.

The full Quick insights capabilities are only generally available in the Power BI service for import mode datasets without RLS applied and with dashboard tiles reflecting those datasets. However, the essential capabilities of Quick insights are also available in Power BI Desktop as the Analyze feature.

In Figure 8.23, Quick insights has been executed against the Chapter06_Import dataset in the Power BI service:

Figure 8.23: Quick insights for a dataset in the Power BI service

To execute Quick insights against an entire dataset, see the...

Mobile-optimized report pages

A critical use case for many reports is access from smaller form factor devices such as phones and tablets running the iOS and Android mobile operating systems. A report that is perfectly designed for a laptop or PC monitor may be difficult to use on a tablet or mobile device, thus depriving business users of advanced analytical insights while they are traveling or otherwise on the go.

To account for multiple form factors, including both small- and large-screen phones, report authors can create mobile-optimized reports via the Phone Layout view in Power BI Desktop.

In Figure 8.26, the Mobile layout of a report page in Power BI Desktop is accessed via the View tab:

Figure 8.26: Mobile layout in the View tab

From the Mobile layout view, the visuals created and formatted for the report page can be arranged and sized on a mobile layout grid. In Figure 8.27, the two KPI and card visuals included in the preceding image from the Report view...

Summary

This chapter reviewed many advanced analytical and visualization features that are available to deliver powerful and compelling report content. This included the use of more complex visuals such as the AI visuals, R and Python visuals, the ArcGIS Maps for Power BI visual, and custom visuals. Additionally, the analytical potential of animation via the standard scatter chart visual and custom Pulse chart visual was explored. Finally, the advanced analytical capabilities of the Analytics pane and the Quick insights/Analyze feature were presented.

The next chapter utilizes the report visualizations and design patterns described in this chapter as well as the previous two chapters to create Power BI dashboards.

This includes simple single dashboard projects and more elaborate multi-dashboard architectures, representing different levels of detail. Although some users may only view or interact with Power BI via dashboards, the quality and sustainability of this content,...