Machine learning, embedded analytics and big data march ahead in 2014

A lire sur: http://searchcio.techtarget.com/opinion/Machine-learning-embedded-analytics-and-big-data-march-ahead-in-2014

Nicole Laskowski, Senior News Writer

Will we look back on 2014 as the year that marked a new era in business? According to faculty at the International Institute for Analytics (IIA), advances in big data, along with machine learning and embedded analytics, will drive new products, reinvent old business processes, and quite possibly mean a lot more work for lawyers.

The Portland, Ore.-based advisory and research analytics organization released nine official predictions for the new year, but based on a recent webinar where faculty members discussed what they see for 2014, the number was closer to 12, by my count. Here they are:

1. Businesses will continue to embrace machine learning and automation. Businesses will need to lean on machine learning and automation if they hope to juggle the incredible volume of data they're seeing. Some businesses, for example, are producing thousands of models -- sometimes at the SKU level. "When you scale that in the hundreds of thousands, machine learning becomes an attractive technology,"Sanjeev Kumar, analytics product manager for Dell's enterprise IT organization, said. "It can continually learn and dig through big data."

2. Businesses will search for the right balance between human and machine capability and judgment. Organizations are getting better at (and more comfortable with) automating business decisions, but that doesn't mean every decision should be automated. "If you take the human too far out of the equation, they lose touch with the work process and their skills can atrophy," said Robert Morison, co-author ofAnalytics At Work: Smarter Decisions, Better Results. And when something unusual occurs or systems are suddenly unavailable, employees won't know what to do. The big automation challenge in 2014 will be finding the right balance, he said.

3. Data scientists will drive new products and services. Evidence for the prediction is everywhere, according to Tom Davenport, co-founder and research director of the IIA and president's distinguished professor of IT and management at Babson College. Here are a few examples: Jawbone, a maker of wearable electronics, hired Monica Rogati last summer. She's a former LinkedIn Corp. data scientist who also headed up the social media company's data products department. Intuit Inc., a financial software company, acquired the small consulting firm Level Up Analytics in October, which is run by data scientists. Gaming company Electronic Arts Inc. is using data to "augment traditional gaming with analytics that tell people how to play games more effectively," Davenport said.

4. Companies get beyond big data hype. As Kumar put it: "2014 will usher in an era where taking stock of what [businesses have] done, where they've gotten, and what value big data and their initiatives have actually brought is going to be high on the agenda."

Rather than keep big data separate, CIOs will need to figure out how to "combine it with other data and feed it into business processes the way they have with traditional data," Bill Franks, chief analytics officer for Teradata Corp., said.

5. Big structured data will make a name for itself. Morison is keeping his eyes on what he called the world of big structured data. "Companies are going to come to terms with big structured data, and it's going to unleash a whole new round of process improvement," he said. General Electric (GE) calls it the "Industrial Internet," and, to prove his point that big structured data is a force to be reckoned with, Morison pointed to the GE sensors from one gas turbine, which generate more data per day than all of Twitter. "That data can be used … [in] product design, product maintenance, management of maintenance facilities" and on and on, he said.

6. The analytics profession continues to mature. Expect to see more analytics teams formed and, even better, more analytics professionals embedded into the line of business. Doing so can create collaborative opportunities and sets the organization on a path to build a center of excellence. "That's the next step, to pull together disparate, business-embedded analytics professionals," said Greta Roberts, CEO ofTalent Analytics Corp.

7. Vendors and users will focus on operationalizing and managing models. "The focus is shifting from classical model development to self-learning/machine-learning model development," said Omer Sohail, head of Deloitte's analytics and information management practice for financial services in the U.S. These more sophisticated models are embedded as close as possible to where decisions are made, with IT focusing less on data and more on metadata and model management. Vendors that supply software and services to build those models are cognizant of the trend, too, and are paying attention to what Dell's Kumar called the model's lifecycle or how models can be deployed, managed and even refreshed so they remain up to date.

8. The adoption of Analytics as a Service will accelerate. With the growing demand for analytics and the continuing talent (and infrastructure) shortage, Morison said businesses will turn to Analytics as a Service products to fill the void. "Factors like speed, cost, capacity and what I'll call 'experimentability' -- all of these things are driving adoption."

PREVIOUSLY ON THE DATA MILL

Add semantic analysis to ward off big data/bad analytics syndrome

Big data tech and bitcoin at MIT VC Conference

Analytics 3.0 -- the old guard masters how to build data products

9. Facial recognition data will play a bigger role. Facial recognition technology isn't just for people. According to Davenport, one pet store wants to recognize pets using the technology (apparently there's an app for that) so they can be greeted by name.

10. Wearable device data will be incorporated into predictive analytics. Activity devices such as those fromFitbit will collect more metrics, which could lend itself to interesting -- and personalized -- predictive analytics, Franks said. Wearable device data could provide insight on everything from the best time to eat a snack to optimize your attention span, to managing stress through exercise.

11. Businesses will embrace visualizations for all types of data. Bubble charts will help to show movement over time just like Hans Rosling did to illustrate disease in the developing world during his TED Talk back in 2006, Davenport said. Organizations will also begin experimenting with what industry analyst Richard Hackathorn callsimmersive intelligence, according to Franks. "It's experimenting with the idea of living data sets -- where you're in a 3-D world," he said. That gives data and non-data folks alike the chance to literally walk through and interact with their data.

12. Businesses will consider adding "lawyer" to their big data lineup? Not every IIA prediction made the list. Here's one that ended up on the cutting room floor: "I suggested a prediction about companies needing to hire lawyers for their big data teams [to ensure] they actually own the data to a sufficient degree so that they can act on it," Davenport said, "but my colleagues were not impressed by that prediction."