IoT Needs Analytics to be Relevant

Three important ways analytics can be an integral part of the Internet of things.

By Bob Potter, Rocket Software

The Internet of things (IoT) has become a major topic in the tech world. Although the concept (but not the term) has been around for decades, recent innovations have made it possible for diverse devices to connect with each other and to centralized IT systems. What does that have to do with analytics? In a word: everything.

The IoT story is primarily a business intelligence story. We are about to enter an era of unprecedented data generation, which means that organizations are going to need to be able to wade through massive amounts of information, make sense of it, and then develop actions based on this knowledge. Connecting 50,000 parking meters together is a neat trick, but actually generating business value requires collecting the data, sorting it, figuring out what matters, and then making decisions based on the relevant information. In other words, we must use analytics to transform data into usable information.

IoT Goes Mainstream

The size of the IoT market makes it difficult to ignore. Cisco IBSG predicts there will be 25 billion devices and machines connected to the Internet by 2015 and 50 billion by 2020. McKinsey reports that, "We estimate the potential economic impact of the Internet of Things to be $2.7 trillion to $6.2 trillion per year by 2025 through use in a half-dozen major applications that we have sized." IBM recently announced a $3.5 billion investment in IoT. Those numbers are staggering and indicate that the need for technology (both hardware and software) to support the massive explosion of devices and data will grow rapidly

Organizations are already getting overwhelmed by the sheer amount of information being created by everything from parking meters to fitness monitors to GPS devices. There's a huge leap between being able to collect fitness data from 15 million joggers and delivering useful applications such as recommended running trails in the Boston metropolitan area.

Personally, I don't get too excited about analyzing data from wearable devices for joggers, but I acknowledge there is a lot of money to be made in this space. I get excited about IoT predictive analytics for patients hooked up to medical devices (both in the hospital and the home) as well as laboratory and clinical medical devices facilitating rapid diagnoses and preventive care for people in medical need. These devices can stream critical data to centralized analytics systems to determine what is going on with the patient compared to his/her peer group. Early detection of catastrophic illness (and potential extension of life expectancy) is made possible by leveraging machines doing the diagnosis instead of the subjective judgment of a doctor or lengthy trial-and-error treatments.

Another IoT use case that "floats my boat" is energy conservation management. Electricity meters have been feeding data to billing systems for decades, but now the meters, the HVAC controllers, and building sensors can work together with public meteorological data to predict and prevent utility brownouts and peak energy consumption at high rates. It also puts businesses in the position to take advantage of demand/response events and save on their energy costs.

Technology, such as Apache Spark, facilitates fast data processing and event streaming to handle real-time data feeds. The latest data visualization products, running in the cloud, allow for users to filter through very large data sets instantaneously and swiftly determine the best course of action - whether it is preventing a loss of life or a factory energy outage.

Think Beyond the Sensor

What can the community of data scientists and analysts do to bring order to chaos? As a first step, organizations need to look to analytics software to quickly analyze and use the data generated by connected machines and devices. We need to get the IoT experts to start thinking about using information, not just collecting it. In the real world, that can take several forms, but here are the three main ways analytics can be an integral part of IoT:

  • Define the parameters that matter. Data discovery tools are good at establishing the right KPIs and filtering out data that doesn't matter.

  • Avoid traditional batch-oriented analytics systems that are great at reporting what already happened. Move to real-time analytics systems based on Spark and/or Hadoop for IoT-generated data.

  • Once the data processing infrastructure is established, use visualization tools that are able to handle very large data sets and can run in the cloud so you can dynamically size analytics capacity based on the specific job at hand.

IoT has received a great deal of attention and traction, but we need to remember that it's still in its infancy. Everyone has a sense that they need to be on the bandwagon, but no one is really sure where it's going. By positioning analytics as the perfect companion to transform Internet of Things data into actionable insight, together we can make analytics an integral element of a successful IoT strategy.

Bob Potter is senior vice president and general manager of Rocket Software's business information/analytics business unit. He has spent 33 years in the software industry with start-ups, mid-size, and large public companies with a focus on BI and data analytics. You can contact him at

TDWI Membership

Get immediate access to training discounts, video library, BI Teams, Skills, Budget Report, and more

Individual, Student, & Team memberships available.