The Microsoft Business Analytics Stack Summary
Microsoft offers a range of business analytics tools and platforms that are unequalled in their breadth and diversity. They address business intelligence (BI), predictive analytics, mobile BI, multidimensional analysis, statistical analysis, data warehousing, data exploration, data discovery and visualization, big data and collaboration. In fact Microsoft has consistently ranked highly in Gartner Group’s rating of business intelligence platforms, and 2014 and 2015 saw acquisitions and several new products.
Microsoft is making a bet on the future, and it’s a bet that will have a big payoff. Many of the pieces are in place – the market just needs to catch up.
Business Intelligence
Starting with the most widely used tools, millions of business users employ Excel to analyze their data. Excel consistently gets bad publicity for the islands of analysis it can produce, but at the end of the day users find it relatively easy to use and powerful enough to do many forms of analysis. Power BI was based on extensions to Excel, with Power Pivot supporting the creation and manipulation of pivot tables, Power View allowing users to explore and visualize data, and Power Query for joining and preparing data for analysis. The introduction of Power BI Desktop this year gives users a more visual interface, somewhat similar to the new breed of data visualization tools such as Tableau and Qlik Sense. Many other businesses use SQL Server as their database and develop reports using SQL Server Reporting Services (SSRS) to create paginated reports (invoices, purchase orders and so on). This is not exactly the easiest tool to use for reporting purposes, but various alternatives are available. SQL Server Analysis Services (SSAS) adds multidimensional analysis, which can be exploited within Excel, but also within Power BI Desktop (which is free to download). It also includes data mining tools to create predictive models.
While the IT industry has recently seen a feeding frenzy for data visualization tools, the demand for operational reporting remains as strong as ever, and Microsoft now classifies four types of reporting:
- Paginated reports require precise formatting and tend to be run on some sort of schedule. The demand for paginated reports will always be with us and Microsoft satisfies this need with SQL Server Report Builder or SQL Server Data Tools.
- Interactive reports can be built with Power BI Desktop, and these allow users to drill down to detail in an interactive manner. Reports can also be filtered and sorted as required.
- Mobile reports are facilitated through Datazen. Microsoft acquired Datazen in 2015, giving it a platform for mobile BI, where dashboards and KPIs can be viewed on mobile devices via native apps.
- Analytical reports are primarily the domain of Excel and particularly the use of multidimensional cubes created via SSAS.
Power BI Desktop, in conjunction with cloud based Power BI Service provide a contemporary platform for self-service data discovery, exploration and visualization.
Data Mining, Predictive Analytics and Machine Learning
Business intelligence concerns itself with a look in the rear-view-mirror – with historical reporting, and dashboards that show the current status of operations. Data mining uses machine learning algorithms to find patterns in data which might support more accurate decisions in the future – who should be given a bank loan for example. This form of analytics is known as predictive analytics – predicting the outcome of future events, or at least improving the accuracy of associated decisions.
The data mining support within SSAS offers a surprisingly good platform for creating simple predictive models, but Microsoft’s positioning in this space all changed when it acquired Revolution Analytics – a commercial version of the open source analytics platform R. This support for R has been broadened with the introduction in SQL Server 2016 of SQL Server R Services. Enterprise scale statistical analysis and data mining is now possible, and with SQL Server 2016 can be deployed easily in a production environment, although Revolution Analytics also provides enterprise scale deployment options.
Also in 2015 Microsoft announced its Azure Machine Learning cloud based service. This has become part of the Cortana Analytics Suite, and provides developers of predictive models with the tools and infrastructure they need. It is also a market place where developers can publish and consume machine learning solutions. Microsoft has consolidated Azure Machine Learning into Cortana Analytics to provide a combined big data and advanced analytics suite.
Cortana Analytics Suite
A single phrase characterizes where Microsoft is going with its analytics platforms, and that is “real-time’”. Cortana embraces big data, machine learning and analytics, and data visualization in a managed cloud platform. It contains elements such as Azure Stream Analytics to handle real-time data flows, which in turn can be connected to real-time dashboards developed in Power BI. The data stores include Azure SQL Data Warehouse and Azure Data Lake. This latter uses big data technologies to create a massive data storage facility.
Data
While SQL Server 2016 is obviously a significant move forward for the product, it is the Azure cloud based databases that have seen most development. These include:
- Azure SQL Server in the cloud (a hosted version of SQL Server on Azure)
- Azure SQL Database – native to the cloud and offered as a service.
- Azure SQL Data Warehouse – large scale data warehouse with support for Transact-SQL (an extended form of SQL)
- Azure HDInsight – a managed Hadoop, Spark, HBase and Storm environment
- Data Lake – using big data technologies (Spark, Hive, HBase and Storm)
Microsoft is also providing tools and infrastructure for data in motion, as well as stationary data held in databases. Azure Stream Analytics is central to this. It preempts arrival of the Internet of Things (IoT) and facilitates real-time analysis of high volumes of streaming data. It also connects to Power BI dashboards for real-time display of data.
Conclusion
There is an awful lot here, and this ‘awful lot’ represents a major investment for Microsoft. The company is serious about business analytics, and provided a business is happy buying in to such a broad set of technology from a single supplier, they will find their analytics needs can be satisfied – no matter the size of the business. And obviously many third party suppliers of analytics technology plug into the Microsoft ecosystem, and so the total breadth of capability is really quite unique.
Microsoft is anticipating the future with its technology stack, and as far as I am concerned they’ve pretty well got it right. The missing component is a decent prescriptive analytics component – analytics which concerns itself with the optimal deployment of resources. Excel does provide a solver, but it is inadequate for large scale optimization needs. So perhaps we’ll see another acquisition soon!