Analytics for IoT (the Internet of Things) is a hot topic within the ever-increasing Analytics discussions now happening in multiple industries. “What’s the difference between Real-Time IoT Analytics and IoT Data-Mining Analytics?” is a question we frequently hear. This post offers a high-level perspective on these two types of IoT Analytics.
1. IoT Data-Mining Analytics is the traditional way of conducting analytics. It focuses on the analysis and measurement of data mined from various IoT sources with the intent to identify past activities, situations, and their implications. The goal of IoT Data Mining Analytics is to understand a target environment as it was when the data was collected.
2. Real-Time IoT Analytics is a new trend in IoT analytics focused on understanding the data as it is gathered. It focuses on moment-by-moment analysis and measurement of data from a live, or “real-time”, IoT environment. It provides immediate feedback on the current environment, activities, and situation and reveals the implications of those events in the environment. The goal of Real-Time IoT Analytics is to understand a target environment as it happens.
Real-time data exists throughout IoT environments — IoT devices and sensors inside equipment are continually generating information about their status, location, performance, and so-on. IoT Analytics must therefore be processed in real-time from the edge to the cloud/data centre.
As you likely guessed, both types of IoT Analytics have their purpose. Possessing historical data and being able to mine and analyze recorded information is valuable in several situations: identifying trends, predicting future trends, and understanding “what happened?”.
Historical data does not provide live and immediate visibility into the impact that a changing variable has in the environment. Although data mining is essential, it analyzes the past, not the present.
In today’s critical IoT environment, it is not acceptable to wait for hours, days, or weeks to know if a change has a positive or adverse effect.
Real-time analytics is especially important when the environment in question is mission-critical, and machine or human actions can adversely affect the environment, health, safety, and production.
Real-time analytics produce results in seconds, allowing users, systems, machines, and consumers in the internal or external IoT environment to visualize immediately, as it happens, the impact of a dynamic change. Real-time IoT Analytics best meet the ever-growing need for instantaneous, actionable information most mission–critical operations need to deliver safety, reliability, and operational effectiveness.
Data-mining and real-time IoT analytics are complementary. To benefit from immediate, actionable information and feedback, IoT analytics must be real-time. To properly document the results of change and to go back in time to examine impacts and implications, data mining is essential.
In a mission–critical environment, IoT analytics must be real-time, and the data should be stored to allow for historical data mining to audit and review changes in the IoT environment in the future, such as for post-mortem investigation. That said, real-time IoT Analytics has the potential to avoid critical or fatal events in the first place.