Smart devices give us data but they don’t automatically make us smarter

January 6, 2016

Posted by: George Malim

Matt Davies, Splunk

Unless we do the right things with data from smart devices we won’t become smarter, writes Matt Davies the head of marketing for EMEA at Splunk. The flow of information from sensors within machinery, devices and any other products that have the ability to create data – which could be anything – opens the door to becoming more connected and operationally intelligent.

However, having the tools to do the job is just the beginning – you see bad workmen blaming them all the time. Good workmen combine them into a toolbox, of sorts, to see all the data in one place and how it fits together.

There are really three steps to becoming smarter in the Internet of Things:

  1. Collect and analyse

You need to pull the data. Installing sensors is great for bigger pieces of machinery, planes, trains and automobiles, but what about smaller devices? Watches, for example. If they can connect to the internet, you can pull data through http. This solves the need for adding sensors onto something an employee might own themselves, or is simply too small.

  1. Correlate across sources

Once you have pulled your data, don’t let it sit in its own database or silo, away from all the other data you’re collecting. Structure your data layer so that you can map different sources on top of one another and run cross-queries. The ability to really explore your data and be inquisitive is key here.

  1. Really use the data

This means analyse, visualise and democratise the data. Allow more people to easily access it. The secret is not to make your data inflexible by applying a rigid structure to it. By adopting a schema on the fly approach, the questions you can ask of IoT data can change and evolve very quickly. This means that different users can ask their own, very varied questions of it. Natural interrogation of the data will start to help you become more proactive and predictive – spotting the signs of upcoming failure, so you can conduct preventative maintenance, for example.

Industries drive a lot of machine data and the way that’s happening is changing. What the ‘server’ looks like today is very different to 20 years ago. It’s not a box in a data centre anymore; it’s a car, it’s a train, it’s a factory milling machine. With all of these new, but highly valuable sources, it’s important to bring ease of use and time to value to big data.

There are early adopters for this approach throughout industry. Zebra collects and analyses machine data taken from the manufacturing floor. This lets it spot problems before they happen and gain insights that drive improvements.

But the benefits don’t stop at the factory gates. When there are devices out in the field, businesses can get real time insight that supports critical business decisions. New York Airbrake (NYAB) monitors the output from multiple sensors on the trains it maintains, even correlating against external factors, such as fuel prices to get more insight into cost. The outcome is that NYAB gives its customers a US$1bn return.

But it’s not just about what’s already happened. What will happen next? Combing data sources for recent and current trends that predict what will occur next give businesses a step up. Gatwick Airport in the UK, for example, looks at the passenger flow within the airport, but also the road traffic that’s incoming to predict passenger volumes four hours in advance. They have moved from “how did we do?” to “how are we doing?” and can now answer “how will we do?”

Analysing, interrogating and democratising data opens up all sorts of opportunities, many of which the business may not have dreamed of. We recently saw Deutsche Bahn host a 24-hour hackathon. In fact, a Splunk team won, by finding an innovative new solution to identify track defects. Not only does democratising data lead to innovation, it leads to a more inquisitive culture around data, which will keep producing innovations.

It’s not just about having connected devices, it’s about having a data strategy behind them to make them intelligent. Getting stuck into the data quickly, easily and accurately means that your business can take advantage of these exciting new sources of data and come out on top.