3 Reasons Why IoT is Key to Solving Your “Small” Data Problems
IoT provides a cost-effective way to gain access to real-time data. While data is presumptively valuable upon analysis, the sheer magnitude of data that is potentially available, and that can be accumulated over time, is daunting. Many have heard of Big Data, but such a term can represent a misnomer to those actually leveraging real-time data to produce true business value.
It is simple to conflate the ideas of IoT and Big Data. IoT can certainly produce Big Data, since many envision billions of connected devices continually pumping massive amounts of data into a cloud-hosted data lake. What then? The specter of being backed into a corner with such a Big Data problem rightly raises the question of whether an IoT solution should even be pursued.
The Big Data problem is not the only problem that IoT solves. IoT is perfectly suited to solve your “Small” Data problem.
Big Data has often been associated with data mining, artificial intelligence, machine learning, predictive analytics, and other processing-intensive exercises focused on deriving insights from patterns hidden within the large data set. In other words, those insights may not be readily apparent from the data at face value without drilling deeper into the data. The more historical data available, the more potential there is to derive deep insights from the large pool of data.
“Small” Data, on the other hand, can represent a limited pool of data that provides insights without a processing-intensive exercise. We explore some reasons as to why IoT is key to solving “Small” Data problems.
“Small” Data Fixes Problems Happening Right Now
A simple example of “Small” Data is data that tells you what is happening right now. Real-time data, for example, can tell you about what a device, machine, or system is currently doing. Real-time visibility into current machine operation provides insights into actual malfunctions that are impacting operations. Wouldn’t it be helpful to know when a device, machine, or system stops working?
In a simple example, a sump pump that is not operating when it ordinarily should be operating (e.g., during a prolonged rainstorm) will provide an immediate alert to facility management teams of a likely sum pump malfunction. Real-time IoT data that provides visibility to the on/off operation of the sump pump solves an immediate “Small” Data problem. Deep insights through Big Data may be helpful to determine predictive maintenance for the sump pump, but those insights would not be required to solve the most immediate operational problem, a sump pump failure!
In many cases, small amounts of data are sufficient to solve huge operational challenges.
“Small” Data Does Not Require Advanced Analytics
To many, the term “Analytics” often implies advanced metrics and inherent complexity. This perceptional bias is part of the reason why “Analytics” is conflated with Big Data. Big Data certainly leverages Analytics; “Small” Data also does.
Again, many consider Big Data from the perspective of a lot of data. The massive volume of data can be derived from a large historical collection of data from a single machine (Big Data), or recent data from each of a large collection of machines (“Small” Data). For example, Big Data insights can be gained in analyzing patterns from three years of data from a single machine, while “Small” Data insights can be gained in analyzing states and conditions from one week of data from a collection of machines.
“Small” Data can yield simple, yet powerful analytics, such as (a) How many times has a machine been on in the last 24 hours?, (b) What is the longest duty cycle of the machine in the last 24 hours, (c) How much energy did the machine consume on average in the last 24 hours?, and (d) What was the impact of the machine on average in the last 24 hours? Visual inspection of any one or more of these “Small” Data KPIs would provide operational insight into potential problems.
To those Subject Matter Experts (SMEs) (e.g., facility management personnel) that are familiar with the machines, “Small” Data visibility of current and recent machine operation will provide insights upon immediate visual inspection of the “Small” Data.
“Small” Data Can Leverage Existing Infrastructure
IoT can solve problems at different levels of scale, from targeted to comprehensive. When viewed from the lens of “Small” Data, IoT can be leveraged to capture only as much operational data as needed. A comprehensive overhaul of the data collection infrastructure need not be required. Capturing operational data directly from existing equipment will greatly reduce overall project expenditures and maximize return. For example, retrofit IoT solutions can digitize critical HVAC equipment such as cooling towers, chillers, RTUs, AHUs, etc. Operational insights from "Small" Data from the HVAC equipment will go a long way to producing efficiencies and cost savings.
IoT is particularly suited to leverage existing infrastructure to the extent possible to extract the “Small” Data needed in that context. The focus should be on capturing the right sensor data for operational insights, not all possible sensor data to develop a data archive for future advanced analytics. Getting the right sensor data to the right SMEs is more important than solving a data architecture model from an IT department perspective. Don’t let a Big Data mindset derail your IoT project.
More importantly, don’t let your “Small” Data project get Big.
Conclusion
IoT has been prone to hype because the promise and potential of real-time data collection can solve the largest of problems we can envision. Don’t get distracted! IoT delivers crucial “Small” Data that will transform any organization looking to generate operational efficiencies.