Here Comes the Analytics Economy

Monitoring IT assets, the software they are running, their configurations, and their software upgrades is nothing new for IT organizations. This makes up the basic operational duties for data centers. But what if your IT operations are spread out over thousands of remote devices that may be in motion much of the time and out of range of the network?

Managing that type of network and gaining insights via analytics from such a network is a newer challenge. It's what GE Transportation is doing with its network of devices to track train locomotives. GE Transportation leader Garret Fitzgerald joined SAS Chief Marketing Officer Randy Guard on the main stage at the SAS Analytics Experience event in Washington this week to talk about the challenges and benefits that the Internet of Things presents. (SAS is sponsor of the AllAnalytics site.) The use case Fitzgerald discussed is really a classic one for the IoT and analytics at the edge.

That use case is also an example of something else -- what Randy Guard says is the emergence of the analytics economy.

"What really defines the analytics economy is the acceptance and pervasiveness of embedded analytics," he said in his presentation this week. "In this new economy each insight sparks the next insight and then these insights compound, just like our investments."

SAS CMO Randy Guard (left) and GE Transportation leader Garrett Fitzgerald (right).

SAS CMO Randy Guard (left) and GE Transportation leader Garrett Fitzgerald (right).

Guard said that the key to the analytics economy are three things -- data, analytics, and collaboration. And collaboration includes both people and machines.

Fitzgerald spelled out a use case for this new economy by describing GE Transportation's work to create IoT, management, and analytics solutions for its locomotive customers.

Fitzgerald started by sharing some data points. In North America alone there are 26,000 locomotives pulling freight. The longest running locomotive in the world is now over 7 kilometers long and can pull 100,000 tons of freight.

Assets like these locomotives can now be equipped with sensors that generate data. But they travel in and out of range of the data network.

"This is not an environment like your data center where it is temperature and humidity controlled at a fixed location," Fitzgerald said. "So as we think about deploying analytics to the edge, we are not just thinking about the outcomes we are trying to achieve. We are also thinking about how to do it in an effective way."

Fitzgerald said that there are two decision points to consider when it comes to determining whether to run analytics at a centralized data center or at the edge. The most important is data availability.

"Mobile assets are constantly moving in and out of coms," he said. "It just wouldn't be feasible to stream all that analytics off board. So you run your analytics at the edge, make decisions, and then send that alert back."

The second decision point hinges on where you need to take action. If you need action to happen in close to real time, then you need those analytics at the edge.

"We are trying to do the same as automotive, go to drone trains, go to autonomous train operation," Fitzgerald said. "Especially when you are talking about things like autonomous train operations you need the decisions made at the edge. You need the insights acted upon by the assets at the edge."

Another application from these solutions can come in the form of fuel cost analysis leading to cost reductions, according to Fitzgerald. He said that railroads spend between $10 and $14 billion a year to fuel locomotives.

"You can think of how even a 1% improvement in efficiency, driven through IoT and edge analytics, could dramatically impact profitability."

Fitzgerald's use case is just one example of technology that can be applied across many different industries, according to Guard.

"The analytics economy will bring a new era of change," he said. "It's incremental in some cases, but it's also very transformative. It's across the board from technology, to business process, to new aspects of industry that we haven't seen yet."

Jessica Davis, Senior Editor, Enterprise Apps, Informationweek

Jessica Davis has spent a career covering the intersection of business and technology at titles including IDG's Infoworld, Ziff Davis Enterprise's eWeek and Channel Insider, and Penton Technology's MSPmentor. She's passionate about the practical use of business intelligence, predictive analytics, and big data for smarter business and a better world. In her spare time she enjoys playing Minecraft and other video games with her sons. She's also a student and performer of improvisational comedy. Follow her on Twitter: @jessicadavis.

Success Secrets of Top Omnichannel Retailers

It's a tough and changing environment for retailers. Yet some are enjoying continued success during turbulent times. We take a closer look at how they do it.

A2 Radio: Lean Analytics for 2018

Lean Analytics author Alistair Croll joins AllAnalytics radio to talk about how to apply the process for 2018.

Communicating locomotive data
  • 10/2/2017 9:46:00 PM


Michelle asks

I would like to learn more about connecting sensors to dump data once they are in [range]. How does GE get data from these sensors that are mostly disconnected from a hub?

Interesting question. I don't have a precise answer (one might need to contact GE's Garrett Fitzgerald to explain), but I can make some guesses.

A couple of my previous articles for A2 have focused on modern railroads' analytics-based technology:

• US Freight Railroads Roll on Analytics
• Are Railroads Headed for Big-Data Overkill? 

First, sensors could transmit specialized data from train cars or locomotives to wayside receptors, but this would be intermittent and dependent on the location and spacing-frequency of the receptors.

For locomotives, the sensor data could be conveyed via internal wiring (or even wirelessly) within the locomotive to the onboard computer, if it's so equipped to receive such data. From there, the data could be transmitted (perhaps via WiFi) to the internet (in effect, an IoT link). However, this would require the WiFi capability and access to WiFi hotzones or a continuous WiFi Internet link.

Another possibility is to transmit the special data via the positive train control (PTC) system. I explained PTC in the second article listed above, and now it's nearly totally in place. PTC requires a continuous digital communications link along almost all railroad lines.

This link might be adapted to enable the transmission of the specialized locomotive performance data that Fitzgerald mentions.

However, it's the railroads that own and operate the communications technology to facilitate data transmission by PTC, WiFi, or another means. So GE would have to make some arrangement with each railroad operating their locomotives to "piggyback" its own data flow on top of whatever system each particular railroad is using.


there's more
  • 9/30/2017 7:05:12 PM

I would like to learn more about connecting sensors to dump data once they are in rance. How does GE get data from these sensors that are mostly disconnected from a hub?