Power BI updates in 2019 so far: Enterprise-grade improvements, pre-built tools show promise

July 15 2019

Matt Allington

Microsoft has continued a steady stream of new and enhanced features for Power BI in 2019. The changes have broadened its accessibility both to end users and to experts like data scientists, security and governance chiefs, and finance executives.

The audience for Power BI is growing, says Microsoft MVP Matt Allington, and that includes end-users in search of education. Allington, who is based in Sydney, provides Power BI training in Australia and beyond. He spoke with MSDW about some of the recent updates to Power BI and how users with different levels of expertise are adapting.

Over the past two years, the number of people interested in training has increased. [And] there are people who are more end-users rather than professional data analysts. I think that's driven by awareness, and the fact that it's being deployed as a meaningful part of reporting.

While he offers public training for prospective users from companies of any size, many of the trainings are two day, in-house events for larger organizations—or online. When he works in-house, Allington interacts directly with organizational data or standard content to help make examples more relevant.

Building an enterprise-strength solution

Allington identified key changes to Power BI within the past 12 months. He sees Microsoft taking critical steps to turn it into an enterprise-strength solution that competes with traditional players.

Power BI has always been very well structured. Built as self-service tools, enterprise strength rigor is given to people building their own reports. Power BI is a stable and consistent product for repeated and repetitive reporting. [Microsoft is] going back and upscaling that technology so it can also cater for enterprise BI reporting. In the past those companies would use SQL Server Analysis Services, ClickView, Tableau or one of those [systems]. It is changing and has changed so they don't disregard PBI, with its ability to share data and have version control.

[Microsoft] introduced a feature called dataflows. In its simplest form it's the Power Query technology available in the cloud. Dataflows is the ability to effectively build a data warehouse online without having to put in typical data warehousing infrastructure. Although it doesn't replace data warehouses yet it has the potential to. It also brings in the Common Data Model and the Common Data Service and this is where it has a big impact on Dynamics 365. It is the standard way to extract data into dataflows for easy, consistent consumption in Power BI reports.

Looking at Dynamics for building reports in Power BI it was quite difficult 12 months back. Extract into Excel was relatively straightforward but not scalable. With dataflows the end to end process will be completely automated [although it's not quite there yet]. [Accessing customer CRM data without manual intervention] is quite literally infinitely easier. Companies with multiple enterprise systems can use dataflows and the Common Data Service to bring cloud based access to their data.

Not all organizations can afford to have teams of data scientists and Allington added that AI and machine learning capabilities are another major area of growth for Power BI.

Microsoft has been pretty good in this space for a long time with Azure ML, quite a specialist area like Cognitive Services. [Previously you needed people skilled in statistics to] build predictive algorithms and then you could actually run the algorithms to either analyze data in the past or predict what might happen in the future. [Now customers have] standard access to, for example, sentiment analysis…It's really quite amazing in demonstrations the quality of what you can do, without connecting directly to Cognitive Services.

If you are lucky enough to have a data scientist, plug in predictive algorithms in the back-end for general users in Power BI…If you're a telco and a data scientist is trying to predict when a customer might churn based on calling or payment patterns, they can build the algorithm and make it available. Others can consume the model without being an expert.

Another key change is Microsoft's recent addition of paginated reports and certified datasets. According to Allington, until recently some companies hesitated to use Power BI because of needing to print reports on a regular basis. Although it offered great features to "slice and dice data," Power BI was not well optimized for printing until recently.

Now you can set header and footer and data in-between gets paginated. There's always going to be a need for some people to print reports in that way … [Before certified datasets], really only self-service people were well supported for producing reports on a topic. If you were the IT department responsible for deploying business intelligence that infrastructure wasn't robust enough. Now IT can come in and say 'This is the one we endorse.'  Desktop users can then create their own reports from the officially endorsed data.

The path forward

About Eamon McCarthy Earls

As the assistant editor of MSDynamicsWorld.com, Eamon helps to oversee editorial content on the site and supports site management and strategy. He can be reached at eearls@msdynamicsworld.com.

Before joining MSDynamicsWorld.com, Eamon was editor for SearchNetworking.com at TechTarget, where he covered networking technology, IoT, and cybersecurity. He is also the author of multiple books and previously contributed to publications such as the Boston Globe, Milford Daily News, and DefenceWeb.

More about Eamon McCarthy Earls