Thingmonk “In education technology one thing I’d love to see is … making sure every coding club in this country offers hardware as a topic,” said Alexandra Deschamps-Sonsino, giving the opening talk at yesterday’s Thingmonk Internet of Things conference in London’s Old Street district.
On the final day of the three-day conference, located under an East London railway arch, which rumbled every 10 minutes as London Overground services rattled overhead, 80 people crammed themselves into the stifling space to hear what’s hot and what’s not in the Internet of Things (IoT).
Aside from the bizarre “particle internet button” (a polo shirt with a doughnut of LEDs sewn into it) which, we were earnestly told by IBM’s Greg Gorman, tracks Twitter sentiment towards Donald Trump and also tells you the weather in Dallas, Texas on command (it was a proof-of-concept rather than something you’d wear, as he demonstrated by wandering around lunch wearing an ordinary polo), there was plenty for the enterprise-focused IoT aficionado.
Just because you can doesn’t mean you should
Dr Boris Adryan gave a talk on the reality of deploying sensors and running analytics in real-world IoT applications – “cutting through the marketing bullshit”, as he put it. An IoT and data analytics man, Dr Adryan joined Zuelhke Engineering in Frankfurt earlier this year. From his research, he learned that 39 per cent of SMEs are worried about the upfront cost of deploying IoT.
Using the example of a trial of smart parking spaces in the London Borough of Westminster, Dr Adyran demonstrated that some straightforward analytics applied to the initial usage survey data could have reduced the number of deployed sensors by two thirds.
“There’s an air of magic around data and analytics,” he said, adding: “People have unrealistic expectations of why they need analytics,” as he demolished the notion that the only way of doing IoT is to pack every single available surface with sensors. “Our potential customers are not in IT but you may come to a situation where they don’t quite understand what they’re activating.”
Returning to the title of his talk, Dr Adryan then took a healthy pop at the current fad for deep learning: “Even script kiddies now can do deep learning with no idea about how to use this and how to interpret the results. Proper machine learning is about getting the right solution for your problem.”
The Web of Things, or ‘buy my book’
Dom Guinard, CTO of evrythng.com, gave the audience a verbal synopsis of his new book, Building the Web of Things. This was billed as a “step-by-step book teaches you how to use web protocols to connect real-world devices to the web.”
“The Web of Things was coined in 2007 by a couple of researchers and myself,” said Guinard. “We realised the state of the IoT wasn’t great… what the IoT should be about is converging networks into the internet and into the web… Protocols should converge at the application level to integrate with devices.
“We wanted to encourage people to build web-connected devices,” he continued, “and provide an end-to-end methodology to build IoT products.” In short, rather than have the plethora of competing standards and technologies that still float around the IoT world to this day, Guinard’s vision is to bring it all together in the same way that Tim Berners-Lee devising the World Wide Web opened up the Internet to non-specialist consumers.
“I think what Node[.js] did was brought the hipsters to the devices,” he added. “It democratised the whole space. I’m not sure I’d run Node on a resource constrained device in production but it allows you to prototype quickly. You see devices pop up that support node very, very well – they even want to cram in Node into their microcontrollers.”
Data scientist Yodit Stanton talked about open data for IoT. “How do we create something useful – bridging the gap between makers and people? How do we fundamentally make a change?” she asked the audience. “How do you create the difference? Every deployment should be challenged. How do you make someone’s life easier, different, better for them? Will it make an impact or will it be exactly the same before?”
Her practical example was an air quality sensor deployment at Heathrow Airport, in areas likely to be affected by the proposed third runway. “The data on noise levels, air quality levels, was sent to Parliament … and now they’re stepping back and doing an impact assessment,” she said.
Augmented Reality is more or less analogous to IoT. According to Stanton, every deployment of IoT amounts to an AR exercise. Giving an example of an architect’s practice designing an office building, she spoke about how data from sensors deployed in similar buildings could be used to deicde where the toilets should be in the new building, or where noisy sales teams should be put so they don’t distract the rest of the company.
“What is the point of what we’re doing?” asked Stanton. Going by the constant chatter at breaktimes about everything from solar panels to chipsets to the breadth and depth of IoT communities and meetups, some people may still be in the build-gadgets-in-your-shed phase, but an increasing number are making things with real-world applications – and making money from them, too. ®
Your correspondent was told of possibly the most pointless IoT device of all time over lunch. Apparently, "Tim" (which is his real name but to spare his blushes we won't use his surname) was at an IoT meetup and, on being shown a Bluetooth toothbrush, said, "I'd never buy one but if you gave it to me, I could probably think up a use case." His friends promptly crowdsourced enough cash on the spot to buy it there and then.
Later on, people phoning Tim early in the morning would hear the answerphone message: "Tim is brushing his teeth right now. Based on past performance you should be able to speak to him in about two minutes."