By Przemek Tomczak
Big Data, Machine Learning, and Innovation
Three days in Barcelona for IoT Solutions World Congress 2019 last month demonstrated the level of maturity of IoT technologies and the importance of developing a business strategy to solve a problem rather than focusing on technology alone. Amongst all the buzzwords and acronyms mentioned, from sensors to digital twins, connected vehicles, M2M, 5G, to edge and machine learning, it was the constant references to data, data analytics, and data processing that most struck a chord with me.
Discussions and interests have shifted from our fascination with predicting how many and what types of devices will be connected over the next 10 to 30 years to how we can analyze, process and utilize the data being generated. There was a general sense that over the next few years, we will see major changes in these areas:
- Fresh business opportunities emerging with on-going challenges in connectivity
- Connectivity becoming a hygiene factor at a commodity level, with the real value in using varied data types and sources to make better decisions
- Rapid adoption of AI, as advancing IoT sensor capabilities combined with the limitations of data being gathered highlight the need to use AI to simplify and accelerate analytical processing
Charting a path for a connected world requires that organizations and individuals make sense of the data being collected and act accordingly. Numerous speakers reflected that “big data is the new oil”. Joe Barkai, author of “The Outcome Economy: How the Industrial Internet of Things is Transforming Everything” and industry analyst pointed out that the value inherent in IoT is being able to “aggregate different types of data to make better decisions.” Kx is well positioned in this new world focusing at the intersection of data and extremely fast analytics for industrial IoT applications as well as our traditional roots in capital markets.
It was evident from the different sessions that many organizations are adopting a more focused and business-oriented approach to IoT and connectivity. It’s all about the data and the primary question should always be focused on what is the problem that I am trying to fix not about what the latest tech can do!
What was interesting was the type of data that was being discussed was predominantly time-series data. Many organizations still view their big data through the prism of “unstructured” architecture and fail to observe that the one constant across all the data is its dependency on time. Being able to capture and incorporate time when ordering and making sense of events can be the main factor in unlocking real cost efficiencies for big data applications. As a case in point, IoT devices generate thousands to millions of data points every second of the day that can be analyzed for predictions, monitoring material failure and making decisions. It is no wonder why we’ve seen time-series databases emerge as the fastest-growing category of databases over the last few years.
Many of my discussions over the three days focused on the challenges and techniques for cost-effectively managing and analyzing time-series data in support of ML-driven applications. ML models need training and re-training based on historical data – increasing the value of historical data. The other challenge and opportunity is bringing ML and analytics closer to where the data is generated – close to the action – whether it is on the factory floor or an oil-well.
My key takeaways from the conference are companies are now actively working on how best to leverage their IoT data whether that is for:
- Situational awareness to gain a rapid view of what’s going on in production processes, logistics or consumer behavior
- Predicting what’s going to happen, whether that is a failure in a machine tool, a process, in purchasing or in revenue streams
- Triggering changes or taking control action automatically or generating an alert referral for human intervention in an automated process
I was struck by the many innovative applications that ranged from predicting failure and optimizing a maintenance program for cars by analyzing vibration levels, to optimizing the configuration of wind turbines in a renewable energy wind-farm, to detecting faults in manufacturing processes to improve yield.
IoT Solutions World Congress Barcelona 27- 29 October 2019 will go down as one of the most impressive events that I’ve attended this year not just for the scale and diversity of the event but also due to the significant growth in the IoT community and the real-world use cases and benefits being realized from IoT across many industries.
I’m already looking forward to IoT Solutions World Congress 2020!
Przemek Tomczak is Senior Vice-President of Internet of Things and Utilities at Kx. For over twenty five years, Kx has been providing the world’s fastest database technology and business intelligence solutions for high velocity and large data sets. Previously, Przemek held senior roles at the Independent Electricity System Operator in Ontario, Canada and top-tier consulting firms and systems integrators. Przemek also has a CPA, CISA and has a background in business, technology, and risk management.