This is part one in a four part series. Other parts in the series can be found here:
Watch this video to learn more about the Future Analytics: The Future of Analytics & Big Data (sapserviceshub.com)
As someone who has been working in Analytics for now over a decade and a half, I have seen a good number of changes in that time. From Decisions Support Systems (DSS) and Executive Information Systems (EIS) to Business Intelligence and Analytics, we’ve seen an increasing sophistication in both the practice and the tools. However, we still are essentially doing the same thing: running queries against an internal data store which is then made available to business users through a user-friendly interface in order to make more informed decisions.
I strongly believe, though, we are now at the cusp of a complete and significant change in Analytics. A number of technologies have become available that changes the landscape dramatically, the needs and speed of the business have changed, and the ubiquitous use of mobile devices is rapidly changing how end users expect information to be consumed. The convergence of Big Data, Predictive Analytics, HANA in-memory computing and data storage, new visualization techniques, a move towards apps and integration with data and transactions coming from the SAP Suite is allowing us to rethink what Analytics can be, and what its role is in a particular business process.
Analytics, in the past years, largely focused on internally held data, typically located in a data warehouse or data mart. This type of analysis restricted itself largely to actuals data, captured through operational source systems. A typical example of this would be a sales report, navigated by product hierarchy and geographic area by a certain time frame. Users typically access a dedicated BI platform environment, separate from the applications they use to perform their actual business activities.
There is clearly value in that, but it leaves massive amounts of information off the table. From email to web traffic, from sensor and machine data to social media analysis, from public governmental data to historical weather reports, there is a wealth of material out there that previously wasn’t even looked at. This is what Big Data aims to address.
Once we add Big Data into our scope of analysis, though, we also need to add statistics and predictive capabilities to make appropriate sense out of such analysis and find patterns and signals in the data. In addition, the combination of Big Data analysis with actuals provides a higher value than either on their own (traditional Analytics alone leaves such content out, whereas Big Data analysis without a link to solid actuals data is vulnerable to higher imprecision and preventable errors, as the discussions around Google Flu have pointed out), so we need an integration platform to combine them together. And since the speed of business has increased and the work force is often more mobile, we’d like this platform to be as real-time and on-demand as possible, and be able to bring the insights to end users where they need it in the business process and in a way that is easy for them to consume.
Imagine, for instance, a supply chain management application that allows business users to see in one place their current inventory, expected use of materials in the coming days, suggested order volumes based on expected demand, puts out a bid electronically for new supplies, recommends the most optimal offer based on price, delivery time and cost, and allows the user immediately to place their order. From their desktop, or from their tablet or smart phone.
The technologies that make that possible are enumerated in the following list:
- Big Data
- Predictive Capabilities
- In-memory database (HANA)
- New visualization techniques
- The App revolution
- Recommended and autonomous actions
The combination and addition of these technologies with traditional Analytics is leading to new analytic applications that enrich such Analytics with information that previously wasn’t available or accessible at all, including unstructured data as well as publicly available or acquired data, and sophisticated predictive models that provide a forecast, or predict behavior based on past performance.
The web has raised expectations around visualization of data and often Big Data and predictive analysis don’t lend themselves well to visualization in traditional bar-, line- and pie-charts, so this is posing new challenges we haven’t often had to deal with before. For instance, the data volume may be such that we simply have too many data points to place in a graph to make sense out of or because of elements of uncertainty in Predictive Analytics and therefore needing to express confidence intervals. If we fail to visualize the appropriate uncertainty, we mislead our end users with predictions that seem very definite, when instead there could be a wide variation. In other cases, it may simply be the nature of the data we analyze, for instance networks of relations, or group membership. Finally, you may simply want to provide visualizations unique to the information and its context and therefore design our charts and graphs ourselves rather than rely on an existing charting package or reporting tool, for visualizations that look more like infographics than traditional reports.
This is even making custom business applications possible, designed carefully around a particular business process, rather than requiring users to open multiple applications and look for information in multiple locations to perform their tasks. We now have the tools through Lumira and SAPUI5/Fiori to quickly develop sophisticated dashboards and custom applications in a matter of days that are unique to a specific business area or business process and can pull all of these aspects together. Tablet and smart phone users have for some time now become used to apps with a unique isolated function. Why would we still expect them to logon and access multiple systems to perform their daily tasks?
Of course, we need these applications to be highly responsive and as near real-time as possible, as well, which is where HANA comes in. HANA also functions as integration platform where all data streams are pulled together and predictive models executed.
In coming parts to this series, we will dig deeper into these technologies and where they fit into the future Analytics. In part 2, we’ll discuss Big Data, predictive capabilities and in-memory and how these three make analysis possible that previously was simply not feasible or took too much time, including on-demand predictive analysis through the in-memory data platform. In part 3, we’ll talk about visualization with Lumira, D3.js and SAPUI5/Fiori and the implications of the App revolution, leading to new delivery mechanisms for analytic content. In part 4, finally, we’ll talk about how integration with the Suite (especially when running on SAP HANA) and recommended and autonomous actions could lead to dramatic changes in the way a business operates.
I hope you’ll join me on this journey.
To learn more about how SAP HANA Services can help you throughout your Analytics journey, please visit us online.