Sensemaking In IoT: Defining Which Data Is Important

Businesses and organizations that fully harness the power of IoT have been able to realize the true digital transformation.  One of the biggest challenges that companies face in turning the potential of this technology into a business-impacting reality is how to manage, store, analyze and action the data that IoT systems generate.

How do you define what data is important and what isn’t? With the high volume and variability of IoT data, how do you define the minimum amount of data points that you need for actionable insights?

First, determine which business problem you’re trying to solve.

Fundamental for a successful data strategy is to start with defining the business problem you are trying to solve. Many IoT projects fail because they are discussed solely from the point of view of the technology rather than from the perspective of the problem you are aiming to fix (the use case) together with clearly defining what you are seeking to achieve (the outcome).

Mining outcomes include ensuring the safety of people and equipment, increasing operational uptime and running operations in the most cost-effective manner.

Oil & Gas outcomes include monitoring real-time equipment performance and inventory levels or planning preventative maintenance activities.

Manufacturing outcomes include reducing equipment failure, increasing employee wrench time and making administrative savings that contribute to long-term success.

Forestry outcomes include improving measurement methods, reducing human intervention on field visits and making faster, smarter decisions that improve profitability.

Then, use your business problem to define your minimum data set.

When you have clearly defined your business problem-led IoT use case, you are then in a position to define your minimum data set based on the types of data you are interested in.

Not all types of IoT data are equal and it’s useful to construct your minimum data set in terms of the following categories of increasing complexity.

  • Status data: whether a device is on or off
  • Location data: tracking the movement of an object or person
  • Automation data: like self driving vehicles, in home devices and assembly lines
  • Actionable data: an extension of status data where the data is transformed into easy-to-carry-out instructions for use in prediction and forecasting, for example supply chain management and building energy consumption.

Build your data set based on the categories of data that meet the objective of your IoT project as set by your business objective.

If, for example, you are interested in building occupancy then your minimum data set need only comprise location data.  But if you wanted to apply the occupancy patterns to building controls such as heat and lighting then you also need to include actionable data.

If you avoid feeling overwhelmed by the volume of data that IoT projects can generate and instead focus on building your data strategy based on your business problem, you can decide which categories of data are essential.