How machine learning analytics can accelerate IoT results

Too often, machine learning requires a massive investment of time and terabytes of data before it can deliver meaningful insights. But that doesn’t have to be the case. A well-configured machine learning analytics tool can rapidly provide initial results — a key advantage for developers who can then start using those results to create value.

To understand this dynamic, companies have to start taking a data approach that embraces the fact that what is driving companies in today’s connected world is not data, but insight. And the volume of data being created in today’s connected world is not just what powers this insight…but also blocks you from finding it.

In order to fully take advantage of a predictive world created by this data, companies need a solution that can help them fully understand the baseline data of their business and what is “normal,” and then apply real-time data over that to power both predictive solutions as well as optimizing future performance. Only a tool that can deliver across all three of these stages can maximize your IoT results.

One such tool is ThingWorx Analytics, part of PTC’s expanding portfolio of solutions that enable customers to create, operate, service and experience smart, connected products. ThingWorx is a platform for building Internet of Things (IoT) analytics-driven applications across a wide spectrum of industries including manufacturing, healthcare, and agriculture, to name a few.

ThingWorx Analytics is different than traditional solutions in that it was built for the world of streaming real-time data. With the volume of data that the IoT ecosystems of tomorrow will deliver, the ability to take real-time data and turn it into timely business insights will drive the return on investment. Without this ability to process real-time data, your company will still be stuck in the old model of batch-driven queries running on top of a static data store. 

What makes ThingWorx Analytics perfect for the IoT world of today and tomorrow? At its heart, ThingWorx Analytics is a machine learning tool that turns real-time streaming IoT data into insights. Compared to traditional analytic tools, it provides application and solution developers an easy way to use advanced analytics methods without requiring expert training in data science, complex mathematics or machine learning. Its algorithms automatically learn from data, detect patterns, and build predictive models, and then send that information to nearly any type of app or technology. That, in turn, produces concrete, real-world results.

One ThingWorx customer is Flowserve, a company that makes flow management products. The company’s sophisticated pumps control and protect the flow of materials in critical industrial applications. If a pump malfunctions on a client’s manufacturing site, costs can run into the millions of dollars. To avoid that, Flowserve outfits its water circulation pump with six sensors to collect data on discharge pressure, suction pressure, and more.

By churning through a stream of 30,000 data points per second, ThingWorx Analytics learns the normal operating conditions of the pump. If pump efficiency drops or an impeller is at risk of failing, ThingWorx sends out a notification. Because the pump self-evaluates all the time, nobody has to monitor it, and problems get addressed before serious downtimes occur.

And in the telecommunications industry, Finnish company Elisa, a market leader in mobile and fixed broadband subscriptions, relies on ThingWorx to power its open service and development platform for customers. Business customers use the platform to build their own IoT business solutions to suit their unique objectives.  

By implementing ThingWorx technology, Elisa can offer those customers seamless connectivity and management of business systems, people, and remote devices. The telco now offers a comprehensive set of IoT services, including data analytics, product development, and device management.  

As the concept of IoT becomes more familiar, how companies can derive differentiated value from it becomes crucial. The real value of IoT lies not in connecting devices (although that is important) but in analytics — specifically, the kinds of powerful analytics that you can derive actionable insights without expertise in complex analytics.  

For more information on IoT analytics, detailed case studies, and more information on getting up to speed with IoT data, please visit

This article was produced in partnership with ThingWorx.

Facebook Comments