Blog IoT Tech Day 2016

The role of data science in this is clear. The sensors involved in IoT will create huge numbers of data streams that cannot be processed using conventional methods. Deriving intelligence and defining clear actions from the data will only be possible after sophisticated predictive models have been built – an area of expertise that lies firmly in the domain of data science.

Talks

The following is a brief summary of some of the more interesting talks that I attended:

1. Unleashing the Data of your Fitness Tracker

The very first talk, early in the morning, was given by Bas Knopper, who in his spare time set out to access the data collected by his FitBit. It was a nice little hobby project that showed you in a very hands-on way how to get the data from FitBit servers using RESTful API calls to their servers, and then process them using Java and Spring as the back-end and AngularJS as the front-end to show a website displaying the personal best results.

2. A Gopher in my Raspberry Pi

Go is a fast, statically typed programming language originally developed by Google. It is gaining momentum due to its high efficiency in both computation and memory, because it compiles to native machine code. Despite this, it is still a rather high-level language with clean constructs and high expressivity, which looks more like Python than some compiled languages such as C++.

The main idea of the talk was to show the effectiveness of using this language on a Raspberry Pi, a microcontroller often used by IoT techies as a basic computing device. This device has many possibilities as a micro PC in controlling and networking IoT solutions, but it has limited resources, making Go especially useful for this device.

Harry de Boer, who gave the talk, showed us how he used Go to measure the moisture content of the soil of one of his houseplants and, if necessary, water it. He then demonstrated how he could quickly create a simple HTTP server to display the water content in real time.

3. IoT Academy

One of the last talks I attended was delivered by Niels Stamhuis on behalf of the IoT Academy, which is a collaboration between KNP, RDM Makerspace and various other partners. He quickly and smoothly presented a number of interesting concepts and projects they are working on.

An interesting aspect of most of their work is how they employ rapid prototyping techniques to get a proof of concept done quickly and then start thinking about a fully-fledged solution, called the proof of solution. Some of the technologies they use for this are LittleBits – which are simple sensors, combiners and actuators that can be clicked together to create simple IoT systems – and Node-RED, which can be used to visually create IoT networks.

4. RIVM

In the final talk I attended the RIVM, the Dutch institute for public health and environment, discussed its involvement with IoT. The continuing drop in sensor prices will make it possible for governmental agencies to better track the quality of air in the Netherlands.

The speaker discussed various sensors, such as those that detect nitrogen dioxide, carbon monoxide, and fine and ultra-fine particles. It seems that, in general, this field profits much less from declining sensor prices as the sensors required for detecting fine and ultra-fine particles are still very expensive A sufficiently accurate sensor costs thousands of dollars. The primary difficulty is that these tiny particles are very hard to detect. The situation is somewhat better in the case of nitrogen dioxide and carbon monoxide, as one can get accurate sensors for several hundred dollars.

The prohibitive cost means that it is still difficult to do fine-grained air quality measurements across the country. The RIVM hopes that these prices will drop further, but understands that this decline is much slower than the decline in the prices of other, simpler sensors.

Conclusion

Most of the talks at the event were technologically sophisticated, well presented and not too much of a sales pitch for products or services. Unfortunately, there were also somewhat less interesting talks. Some of the keynote speeches were overly corporate, for example, filled with hot air rather than anything of substance. And some keynote speakers spent too much time promoting their products rather than focusing on the technology.

Nevertheless, overall, the talks were very interesting. Paired with the networking opportunities, it made the event a success and really opened my eyes to the interesting world of IoT.

Over auteur Gerard Simons

“Sinds augustus 2015 werk ik bij Qualogy als Junior Data Scientist. Ik analyseer en structureer data, om daarmee statistische modellen te ontwikkelen”

Meer posts van Gerard Simons