Internet of What Exactly?
Internet of Things — a buzzword keeping everyone at high alert, is not a new concept despite quite recent rise into fame. It was back in the ’80s that science department at Carnegie-Mellon University in Pennsylvania installed micro-switches in the Coke vending machine to see bottles count on the PDP-10 departmental computer. Then came a toaster, which could be turned on and off over the Internet and that was even before the first web page was published. Since then, we saw radio-frequency identification, barcodes, near field communication, digital watermarking and QR codes applied to not-yet-smart products (to the point that they became a meme).

There is still a lot of talk and little action on consumer side of things, whilst business implementation spreads apace through manufacturing, energy, transportation, agriculture, retail, healthcare, financial services, security and smart cities. It’s impact isn’t easily seen by the general public but it is clear now, that it’s not the China-manufactured “things” what’s interesting about IoT — it’s the data.
IoT applications measure systems operating, technical conditions, monitor failures and investigate their most possible causes to minimize downtime in the future or prevent the ones that are about to happen. The data mined using IoT is descriptive enough to answer business questions of various nature, prevent losses or strengthen the resilience of the process.
Individuals grasping the topic face the challenge of the lack of business intelligence. We learned it the hard way — building the air pollution sensor we had to dive into the meanders of pollution factors, measuring conditions and sensor material limitations for several weeks before making any progress in the project at all. Still, halfway through it, we were at times surprised with unusual findings and setbacks that could have been avoided if we had an air quality engineer in our team. It might seem obvious, but it is only once you encounter a problem such as metal density that you feel the real weight of that principle.
The Data
The IoT is usually a mere toy when it is not combined with Big Data, Cloud Computing and AI. Looking from our perspective, there are countless air quality sensors for both indoor and outdoor use. Some are expensive or quite fancy, other not so much (still supporting though! :)) but the core value of such projects is the data.

IBM saw that and developed the Green Horizons initiative — a global project harnessing the power of IoT and data processing. It uses air quality forecasts to calculate necessary traffic reductions, cut their hidden costs and help state administration anticipate high pollution levels to enforce reduced production at certain factories in endangered areas.
Blue Marine Technologies, SparkLabs IoT program beneficiary, is installing IoT street lamps in Korea to reduce costs and greenhouse gas emissions while improving public safety and providing a wide range of sensor-generated data.
On the other side of the globe, Kansas city is partnering with Cisco and Black & Veatch to fuel smart water initiatives and build extensive system of sensors for leak detection and infrastructure asset management.
Texas will see self-healing Cisco network preventing service disruption in transportation and improving on-time arrivals to over 99 percent.
Data management
IoT data management is hard — with high pace of the development it is challenging to stay ahead, it might be difficult scale and not easy to understand without complex analytics and AI. To get the most out of it, one has to think big too. Sensors often generate thousands of data points every second, so strategic approach to building an architecture that can support it is crucial. Then comes the process of learning the data — the flow, variables, hidden patterns, trends and factors that affect it. A way to meet that is by constructing sandboxes, practice areas where users can experiment with data — ideally with tools, languages, and environments they’re familiar with, as in Gartner’s Heudecker. If there is no clearly defined problem you are trying to solve, maybe looking at a certain business/user behaviour area will give some answers.
Data visualization is also a big leap taken by UI design to tackle the astonishing amounts of information from IoT networks. Still, we think it is best to start small and cheap, look at one variable at a time with a limited proof of the conceptual process. That’s exactly what we had in mind starting with our network of low-cost IoT air quality sensors for public and open-source use. We decided on inexpensive components for continuous testing, pivots and adjustments. The public and open-source character of the project will enable us to deploy numerous devices, check different testing environments and many different variables. We could have gone with just a couple of high-quality, bulletproof sensors and used the average for the entire city but that wouldn’t have been as in depth and we would have lacked the core value of the IoT — the big data. The project will be deployed in the coming months — keep your fingers crossed!

