Alleviating Stress Through Analytics


Analytics give meaningful, actionable insights. Where analytics is going is even more interesting. Today, it’s connected analytics.

By Paul Gillin

It’s inevitable. You have an extremely important meeting today. You look out the window and, yep, those forecasters were correct. High winds and torrential rain. It means one thing: traffic is at a standstill.

You are late for the meeting.

Those stressful situations are becoming a thing of the past.

Austria’s ASFiNAG (Autobahnen-und Schnellstraßen-Finanzierungs-Aktiengesellschaft, if you must know), is helping to ease the pain of traffic when the weather isn’t cooperating.

The state-owned corporation is responsible for maintaining Austria’s 2200 kilometers of roadways. For the past decade the company has been building a fiber-optic network that connects 70,000 sensors and 6500 traffic cameras.

This network manages traffic more efficiently. It also responds to emergencies.  And—yes—adapts to changing weather conditions.

The system lowers speed limits ahead of congestion. This minimizes and even prevents traffic jams.

 It can detect weather events as they are happening. This means road crews can be dispatched to deal with flooding or downed trees.

“Using the data from our networks, we want to provide travel times and even predict traffic issues for a smoother travel experience,” says Bernd Datler, managing director of the tolling company at ASFiNAG.

Smart devices change the rules
What’s helping ASFiNAG do this? It’s analytics.

They give meaningful, actionable insights. Where analytics is going is even more interesting. Today, it’s connected analytics.

Connected analytics is when many variables are in motion. When they are connected, they determine a flow of responsive adjustments.

What else is helping Austria to ease its traffic burden? It’s analytics at the edge. “What’s that?” you might ask.

Analytics at the edge—also known as fog computing—brings all the storage, computing, and networking capabilities of cloud computing “down to earth.” It brings it to the very edge of the network where data is being generated by sensors, cameras, and other devices.

Couple that with the Internet of Things (IoT), and you get the ability to quickly react to events.

In fact, by 2020, there will be 50 billion devices connected to the Internet. Those include machine sensors, thermostats, cameras, and process monitors. All generating constant streams of data.

The big question remains. How do we handle this deluge of data in a way that can yield immediate benefits?

“The amount of data that will be captured by IoT is going to be orders of magnitude larger than we’ve ever seen,” says Jim Hare, a research director at Gartner. “It will require analytics at the edge.”

To understand the need for analytics at the edge, consider an offshore oil rig. It generates between 1 and 2 terabytes of data each day. Using satellite communications, it would take 12 days to transfer one day’s worth of data back to a central data warehouse. That’s clearly too late to use the information to deal with an overheating drill bit.

However, if the data is sorted and analyzed locally, at the edge of the network, the problem can be identified in real time. This means the machine can be shut down and fixed before it causes greater damage.

In oil and gas operations, factories, and smart transportation systems, intelligent-edge devices will be able to make faster, better decisions without human intervention.  

“Machines will talk to each other at millisecond speeds, so if something crashes in a factory, other machines that could be affected can be turned off as quickly as possible,” says David Floyer, chief technical officer of analyst firm Wikibon.

A security camera could be “taught” to recognize irregularities and zoom in for more detail.

Turn off the lights
Energy management is another opportunity for the next generation of analytics. New consumer thermostats learn from the environment, time of year, and the behavior of people in a household. They are capable of consulting central databases for things like weather conditions and typical behavior patterns, then make a decision and adjust locally and automatically.

The same technology applied on a larger scale could have a significant impact on energy expenses. By the way, those expenses make up about 30 percent of the operating costs of a building or factory. The promise of so-called “smart buildings” finally becomes affordable to companies of all sizes.

It’s not only things that are connected. Digitization—enabled by the networked connection of people, process, data, and things—is changing the way data is processed. That means connecting intelligence at the edge with intelligence in the cloud.

Special-purpose devices communicate wirelessly at a local level. Then the cloud is used to consult centralized engines when needed.

Thermostats not only know that a cold front is coming, but draw upon cloud-based analytics engines to adjust temperatures for optimum comfort and worker productivity.

Many of these capabilities haven’t been feasible up to this point because of time or cost.

“Moving data is very expensive, especially over distances,” Floyer says. “The solution is to process as much data as possible locally and transmit only what’s needed.”

Floyer cites video security as an example of the future architecture of analytics.

“If you’re doing video recognition of faces, you’ll see the same people again and again. There’s no reason to send that data over the network,” he says. “But if you see a face you don’t recognize, you’ll want to check that out with a central database.”

Wanted: New approaches
Digital requires re-thinking the way businesses process data. While Digital captures much more data than has ever been possible, only a fraction of that data will be useful.

Knowing what to keep and what to discard is an important factor.

Choosing what data to distribute and how to use it will also become a critical skill. To use the oil rig example, equipment monitors on the platform don’t need to know the entire maintenance history of every connected device.

It needs to know the factors that indicate an impending failure. Those databases will need occasional updates from central sources, as new indicators become known.  

These new approaches will make analytics more complex, but also more valuable to the business.

“CIOs will need a vision that communicates the potential of the IoT to the business and the role that IT can play,” says Hare. “They also need to identify the architectural and infrastructure changes that will be needed to support it.”

Currently, those technologies are still incubating.

Standards for data formats, communications, and security are either in development or scattered across a fragmented ecosystem of competing vendors. “The market is in its infancy,” Hare says. “It’s still the wild west.”

But there are good reasons to settle the frontier.

Those 50 billion connections, by the end of the decade, will have a potential digital value of $29.7 trillion. Savvy executives will want to grab more than their fair share.


Paul Gillin is a writer and senior consultant at B2B social media training firm Profitecture.