The Future of Analytics is Streaming in the Cloud

Michael Vizard Avatar

Updated on:

By Michael Vizard

Business analytics, as it is applied in most organizations, is about to be utterly transformed. Traditionally, analytics have been the realm of a few high priests in an organization looking to divine trends by discovering new patterns in historical data.

But as analytics continues to rapidly evolve, a fundamental shift is starting to occur. Rather than trying to intermittingly optimize a process based on historical data taken from a large number of applications, analytics is getting more prescriptive and making information available in real-time so that it can be directly applied to a specific transaction or event.

A case in point is Emory University Hospital, which is applying streaming analytics developed by IBM to identify changes in physiological patterns to alert clinicians to fluctuations in a patient’s condition in real-time.

“Right now when you check into a hospital a health-care organization knows a lot more about your financial condition in real-time than the clinician knows about your condition,” says Tim Buchman, M.D., Ph.D, director of critical care at Emory University Hospital. “The goal is to turn all the data we have into information that creates more situational awareness for each patient.”

Today most health-care organizations, says Buchman, rely on a trial and error approach that is informed by a retrospective analysis performed across a large number of patients. Buchman says streaming analytics will enable health-care providers to provide the “right care, right now, every time.”

That’s difficult to accomplish without the ability to apply streaming analytics to big data, especially when an intensive care unit (ICU) display is only able to show about six seconds worth of data. Add in the fact that there are usually a half-dozen patients in an ICU at any given time and it becomes impossible for humans to remember and correlate that amount of information. In contrast, a streaming analytics application can make recommendations based on all the data being collected.

“We want our highest paid knowledge workers to focus on the patient, rather than the low-level stuff that can be handled by rote or protocol,” says Buchman.

Predicting Patient Case Flow

The financial implications of analytics being delivered in real-time are just as broad. McKesson, a provider of IT services to the health-care industry, is currently experimenting with Hadoop and the SAP HANA in-memory computing platform to better predict patient case flow for physicians.

According to Gabe Orthous, director of business analytics and product management at McKesson, at a time when physicians are trying to better understand how the Affordable Care Act will affect their income, McKesson wants to be able to work with larger amounts of raw data in real-time to provide physicians with a deeper understanding of what areas of medicine will provide the most rewards at a time when the quality of the service being provided now has a direct impact on profitability.

“We want to be able to run regression analytics every time they pull a data set,” says Orthous. “Then we want to be able to provide the social intelligence to allow them to collaborate with us.”

Moving beyond the health-care industry, Narendra Mulani, senior managing director for Accenture analytics, says this type of analytics capability will be embedded as a service by Accenture within just about every business process regardless of the vertical industry involved.

“We’re going to be delivering insight at the actual point of consumption,” says Mulani. “The analytics will be embedded inside the business function.”

That means instead of having to master a separate user interface for an analytics application, the analytics will essentially be delivered in real-time as a headless service within, for example, the manufacturing application being used to manage the shop floor.

Ultimately, Rajeev Menon, senior vice president of the SAP practice at Wharfedale Technologies, an IT services provider, says the delivery of analytics as a service is how vendors, using SAP enterprise applications evolving on top of the SAP HANA in-memory computing platform, will be able to run transaction processing and analytics applications simultaneously in the same core platform.

“SAP is moving to deliver analytics inside the core business suite,” says Menon. “Moving a data warehouse onto HANA is just the first step.”

Ultimately, the goal is to not only apply analytics in a way that enables CIOs to capture the institutional memory of any given organization, but also to apply that knowledge whenever and wherever it is needed.