Power BI is Microsoft’s cloud based business intelligence platform. What started out as an add-on to Excel and SharePoint has morphed into its own standalone service.
Major changes to the platform have been announced this week that have significant impact on how customers will purchase, interact and deploy this platform.
Power BI vs Power Pivot vs. Power View – Still a Confusing Story
One of the most popular blog posts on this web site is one entitled PowerView vs. PowerPivot vs. Power BI and it describes the key differences between the three products and how they relate to each other.
The new Power BI Preview doesn’t make the explanation easier – in some ways its now more confusing than ever. It’s not clear whether Power Query, Power View and Power Pivot are sticking around, being retired or being absorbed into the new service. It’s also not terribly clear whether the current Excel designers (e.g. the Power Query Add-On and Power View Add-On) are going to be absorbed into the new Power BI designer or continue to be maintained as first class products within the Excel context.
New Visualizations Available only in Power BI Preview
In addition to being HTML 5 based, the new Power BI preview comes with a number of new visualizations to add to your dashboards including:
- Combo Charts
- Filled Maps
- Guages
- Tree Maps
- Funnel Charts
Power BI is no longer tied to Excel and Office 365
One of the key drivers for moving users to Office 365 was that Power BI as a feature set was only available within Office 365. Similarly, you needed Excel 2013 as a key authoring tool to create Power View reports and deploy them to SharePoint online.
While Office 365 and Excel will still work well with the Power BI platform, they are no longer required. Power BI is now independent of both platforms with its own “portal” on PowerBI.com and its own designer for building dashboards. The designer is HTML 5 based which means you could design reports using a basic browser instead of requiring Excel as a desktop client.
Power BI Supports but no Longer depends on Your SQL Server
Power BI is essentially now a pure NoSQL based service. If you look at the APIs, the data sets you can send into Power BI are JSON based and completely abstracted from traditional SQL databases.
While Power BI can pull data from SQL Servers, it also supports a variety of other platforms including non-Microsoft based SAAS services. Expect the list of SAAS services to expand dramatically as preview becomes the full production version.
Expect Deeper Connections to Azure Data Services
Microsoft has been busy announcing a number of new Azure data services including Stream Analytics, Data Factory and updates to HD Insight. Expect to see connections between all these services with Power BI being the visualization layer as data is processed through these new services.
Power BI will Connect to Your On Premise Data Sources
Power BI has had a method for connecting to your on premise data sources for a while now. The new preview includes a similar connector for providing access to your on premises databases. In the new preview, only Analysis Services is currently supported.
Once you have installed the Connector client locally, it broadcasts data from your internal environment up to your Power BI cloud environment.
Power BI for IPAD is coming soon
Microsoft will be releasing a new IPAD app for Power BI that will allow you to render visualizations on your mobile device. It’s still not clear whether there is an Android version coming as well and whether you’ll be able to design reports or just consume them on your tablet.
Pricing Hasn’t Been Updated
It’s not quite clear how you pay for the new Power BI service. In the current model, Power BI is an add-on service to your Office 365 plan or you can purchase it as a standalone subscription.
However, now that the service isn’t directly tied to SharePoint or Excel, this might change considerably. For example, Office Online started out in a similar way with a direct connection to Office 365. However, Office Online is now effectively a freemium service where you can use it with your hotmail account.
Performance is a Key Outstanding Question
One of the key challenges with the current Power BI and Power View platform is performance, in particular with large datasets. Power Pivot is a great solution for doing ad hoc analysis with a couple million records, but if you try to load it up with tens of millions of records in Excel you’ll find it slows down considerably. This is because the PowerPivot model is client cached and dependent on your local machine’s capabilities. In addition, because it is a client cache model, it needs to refresh the data from the original data source and this can take quite a while depending on the location of the data and how much you need to refresh.
It’s not clear whether the new Power BI uses a similar model or whether its connections to underlying data sources are live. For example, if I connect Power BI to my SalesForce.com data, does it create a local cache of the data or is it running queries against my data source directly? Similarly, if I connect to my data warehouse on premise, is Power BI going to cache the entire query in the cloud or will it hit my database each time it needs to run an MDX query?