Project Botticelli

What is Power BI?

10 February 2014 · 1 comment · 3916 views

Happy birthday BI for the next billion users!

Power BI Logo with ExcelHowever sceptical or cautious one might wish to be, there is no denying that Microsoft have attempted something that has potential for greatness: Power BI, which has just been launched, minutes ago, and which is now generally available for purchase.

There is about a billion users of Excel worldwide, according to Microsoft. If anyone has achieved a modicum of business success recently, they have used Excel, at least at some stage. Even if you worked for Apple, who happen to have made a great stab at it with Numbers, chances are you would have still used Excel for all the complex stuff. If you were an astronaut aboard the space shuttle—yes, Excel. Excel is the lingua franca of day-to-day BI, and even if only 10% of the 500+ million Excel users know what it means when they type =A1+B1 that still makes Excel formulas the most successful programming paradigm. Ever.

But the world has moved on. We share and collaborate, and we want better visualisations, and more power! Millions of rows of data about sales of a widget? Power Pivot, an Excel component, will crunch it. Want to animate the rise and fall of the popularity of your star product? Power View at your service, with its cute moving bubbles. Have you just found an excellent way to extract and cleanse useful data from several sources? Power Query will let you share your technique with those who do not know how to write queries or how to cleanse data. Fancy a 3D map showing your sales figures, colourful and informative? Power Map will even let you conduct a fly over, with no pilot’s license required. All of this is now part of Excel 2013, in its Pro Plus edition and in higher Office 365 plans, once you have enabled and/or installed the free add-ins.

But what about sharing and working with others? For better or worse, our offices are no longer bound by buildings’ walls, and we work across distant, virtual teams. Sure, if your company happens to run a recent version of SharePoint server, you are fine and dandy. But what about the smaller and medium companies out there? Or even the big enterprises, who are, perhaps, SharePoint-shy?

Power BI is exactly there for you. Think of it as a pre-built, fully-managed-for-you, SharePoint-like environment that lets you grow an ecosystem of user-created BI.

Start with spreadsheets, power queries, connect to your corporate data, which can reside on-premise, and does not even have to be in the cloud, thanks to the Data Management Gateway. Upload it now to Power BI, and let your colleagues get a new perspective onto what matters to them, thanks to your selection of data, your DAX-enhanced models, and your visualisations. While you still need Excel to create it, you don’t need it anymore to see the results on a laptop, iPad, or on another mobile device, like a Surface, because of HTML5 and a dedicated app (NB, work in progress). Sure, I am concerned about risks of spreadmarts and of varying report quality. Still, user-created data error tends to live within the context of that user’s well-informed work circles, and it becomes an issue only when crossing outside the boundary of that context. A dedicated Power BI Data Steward will help to police that boundary, but more work is needed, especially in larger organisations—this issue does not concern the smaller ones as much.

My favourite part, once you have published and enabled a spreadsheet in Power BI, is Q&A. It is a bit like “Siri for BI” without voice recognition (yet?). You can ask it human questions, like What are daily sales of my purple widgets to wine makers in South Africa, and it will not only parse your question and answer it, but it will even try to show an appropriate chart. Users don’t even need to open the spreadsheet, and they do not need to know which one is the one that has the answer. Think of it as a marriage of the search field, or “spotlight,” that desktops and mobiles have, with an understanding of the semantics hidden in your BI models. I cannot wait till this technology becomes capable of automatically combining external information, and search engine intelligence, with your own data.

Is it all a finished, done-and-dusted product? Way no, not yet! Some of the visualisations, like Power Map, won’t show in Power BI until a future update. Some aspects of mobile device use are still being worked out. But the best part is, that the future of this technology is entirely open-ended, and, in my opinion, simply cannot be finished for good at any point in time. Why? For example, new data sources will have to be added as technologies develop, helping you consume both internal and external data. Sure, traditional BI powerhouses (OLAP and tabular from SSAS, please) will need to catered for, and newer, cloud-based sources, will come along as they become popular—in the meantime, there is OData for everything else. Of course, there will be more visualisations to come. There will be new ways to process it all, and to get more clever intelligence from within it. Sharing to anonymous users, or embedding a Power BI chart on your own web page are musts. No, I do not know Microsoft to-do lists, but I am making an educated guess—and a suggestion.

In the meantime, other than for the somewhat complex licensing scheme, Power BI is ready to get you started on the path to better BI in your organisation, within, and without. Give it a go, today, with the free trial, and then come back here, to learn how to make the most out of it all. Make sure to register for our newsletter, too.

Happy Birthday, Power BI, and congratulations to the entire Microsoft SQL BI, Excel, and SharePoint BI teams. From not-so humble Excel beginnings, this seems like the way to bring meaningful, data-driven, fact-based answers to the next billion users. Best of all, they probably won’t even have to know it used to be called BI. 

Rafal

Comments

Terry Lim · 15 February 2014

this is a great news also for an neophyte on BI, but has started learning & applying powerpivot and basics of data mining too :)