The Data Mill

Gartner urges CIOs to craft analytics strategy for smart devices -- now!

Smart light bulbs from Philips conserve energy and can be controlled by a mobile application; smart contact lenses from Google will soon measure glucose levels in the eye in an effort to battle diabetes; smart textiles will use machine-washable (!) sensors to monitor a person's health and mood.

The Data Mill

"We can change lives," Kurt Schlegel, research vice president for the Stamford, Conn.-based consultancy Gartner Inc., said as he showed off his smart T-shirt from OMsignal during the Gartner Business Intelligence & Analytics Summit's opening keynote. With the click of a button, Schlegel showed an audience of more than 2,000 attendees a live feed of his respiration and heart rate, all measured by sensors woven into the T-shirt and transmitted to his smartphone.

The endorsement of smart devices was actually part of a skit, with Schlegel playing the part of the tech evangelist. He was joined by the skeptic (played by Frank Buytendijk) and the pragmatist (played by Rita Sallam). Together, they delivered an analytics version of Goldilocks and the Three Bears. The moral for CIOs and IT leaders? Establish an analytics strategy for smart devices -- and for the business, in general -- that is neither too hard nor too soft, but rather just right at being decisive, enabling personalization, and providing transparency. Now.

The balancing act CIOs and IT leaders have to perform when juggling these three characteristics is difficult, to say the least. As the world becomes smarter, it also runs the risk of becoming scarier. "Look at LG Smart TV," Buytendijk said. To personalize advertisements on the home screen, LG began collecting (and sharing) viewer information, even when the option was turned off. The CEO later announced the company had gone too far and would release a patch to fix the problem.

"I'm not sure about you, but if a CEO is talking about a patch, and you're in IT, you know you're in deep, deep trouble," Buytendijk said.

Building data science skills from within

Data science skills are still hard to come by, so why not start by building those skills in house. Carol Rozwell, Gartner analyst, provided three tips on how CIOs and IT leaders can do just that.

  • Create centers of excellence so employees within those analytics positions can "leverage best practices and uncover boulders before they trip over them," Rozwell said.
  • Provide career development for junior employees, building up their competency levels to suit the business' needs.
  • "If you do have the advantage of consulting support," use consultants to build data science skills in house, she said.

Watson's architecture

IBM's Watson made a splashy debut on the game show stage, defeating the best Jeopardy! players in the world. Today, IBM is commercializing Watson and building out the partner ecosystem "so it's not IBM determining the applications that will take advantage of this huge engine," Jamie Popkin, managing vice president of Gartner's technical professionals data management strategies team, said at the Gartner BI show.

That huge engine uses natural language processing technology, "both language modeling, as well as machine-learning algorithms," to break down inquiries and generate answers. "There's something like 237 different systems running underneath Watson to do both the language processing and the hypothesis generation, and then the testing, scoring and ranking of the answers," Popkin said.

Previously on The Data Mill

Five tips for a cloud first strategy

Ford's connected car revs up with APIs and external app developers

External social platforms replace the company intranet

Two standout features for Popkin: Like a search engine, Watson displays a weighted set of responses to a question. Some of the responses might not seem like a good fit at first blush, but could suggest a relationship in the data that hadn't been thought of before.

Second, the system recalculates the answer no matter how many times a question gets asked. That means Watson is not storing the answers it provides, recalling them over and over again. Instead, each time a question is asked of it, it starts fresh to include any new information that has been added and indexed into the system.

Say what?!?

"Most of the spending [on big data technology] is coming out of infrastructure: Database management systems are growing faster than BI and analytics -- and so is data integration." -- Dan Sommer, Gartner research director

"Everything is going to become information technology, and we're well along that trajectory already. And if you have enough information, you can create intelligence out of that." -- Ray Kurzweil, author, futurist and director of engineering, Google

"Organizations that treat information as an asset and find ways to sell that data are rewarded significantly -- rewarded in terms of the ratio of the market value to their asset value being 3.5 to 5 times higher than market average." -- Kurt Schlegel, research vice president, Gartner

"We need to adapt to and anticipate use cases in our information infrastructure that don't exist today." --- Frank Buytendijk, research vice president, Gartner

"I'm talking about formalizing analytics as an organization with a seat at the table in its own right. No more competency centers or centers of excellence. It's its own development and delivery organization." -- Jill Dyche, vice president of thought leadership and education, SAS Institute Inc.

Welcome to The Data Mill, a weekly column devoted to all things data. Heard something newsy (or gossipy)? Email me or find me on Twitter at @TT_Nicole.

This was first published in April 2014