IoT and Analytics


It is widely accepted that the Internet of Things (IoT) relates to the concept of ‘widely connected hardware’. Its impact is vast, because any physical object, which has a sensor attached to it, can be turned into a digital data generating machine – rapidly producing more data than most organisations have the capability to report on or analyse in-depth. In terms of BI and analytical insights, the IoT has become a game changer.

So, what is so interesting about the IoT, is the opportunity it presents to the BI and analytics industry. Today, the IoT allows us to tap into new data sources, which makes it possible to track trends using formerly immeasurable methods [1]. This is great in the sense that BI and analytics practitioners now (again) have something new to offer to its clients – that is, if we have the necessary capabilities and we understand ‘what to do’ with all the IoT data.

The impact of data

But let’s look back a bit; remember those meetings that used to be epic because, for example, your client (say a healthcare organisation) did not understand or see how data could affect the day-to-day operations of the business (say, the lives of patients and nurses in the hospital). Back then, data seemed almost intangible. However, now with the advent of IoT, data has become much easier to explain. In our example, using IoT enabled technology, a healthcare specialist can track a patient’s vital signs, and its origins, because now there is a constant stream of recorded data. Surely this translates to real influence.

However, BI and analytics professionals have to step up to the plate in terms of capabilities and skills to harness the power of what modern forms of analytics can do with these new data sources. Unlike traditional sources of data, IoT data streams “arrive” unevenly spaced, which makes it harder to process. In our example, by monitoring this data stream, the healthcare organisation can improve the quality of care provided to patients by tracking their vital measurements transmitted from wearable devices on an on-going basis – from at home before the admission, during the hospital episode, in the rehabilitation unit and then back at home again during the recovery period. This can be applied on a more permanent basis to old age, critically ill and multi-morbid patients – potentially for them even avoiding unnecessary hospitalisations.

More importantly, we can make good use of the IoT’s always connectedness, which in fact “compresses time” as there is almost no lag between an event happening and the availability of the data describing the event. In most cases it is basically instantaneous. In our example, it allows clinicians and nurses to initiate action as soon as any warning signs are encountered [1].

Streaming analytics

We can take this one step further. Applying more advanced forms of streaming analytics, the real-time data feeds can be analysed to predict when a situation that requires intervention may occur – as indicated by a predicted likelihood indicator. This gives clinicians and nurses fore-warning to take preventative action even before a critical situation arises. The lead time gained by such pre-insight can drastically reduce emergency interventions, thereby reducing the application of critical resources (like emergency crews and devices), and it reduces the criticality of the interventions required on the patient, thereby also improving their experience and the overall quality of care.

Noise vs Sense

However, what may slow down the adoption of these types of processes is the fact that when we deal with IoT data, we do not only deal with bits of data, but with massive streams of information generated at previously unheard of speeds. The IoT represents the invention of an online universe of data that’s not limited by physics, distance and time. We have to get used to dealing with the volume and velocity of data generated. There is also a potential sparsity in the IoT data that we have to get used to dealing with – a lot of it may be “regular noise”, but we have to keep monitoring and filtering it for those rare significant data points that may indicate events of potentially crucial importance. A seriously ill patient’s heartbeat that spikes for a second or two may already be a critical indicator, especially taken in context and combination of a number of other key measurements or recent trends.

Concluding remarks

Although I have used healthcare examples here, the application of IoT data and analytics is equally relevant in many other industries too – examples include transport, logistics, security and safety, sport and even recreation and entertainment. In fact, it is applicable in any industry or application where changes can be monitored in real time.

The IoT “phenomenon” not only has the potential to show what modern BI and analytics capabilities can do – but it shows the promise of what insights and resulting actions this new class of data can deliver in organisations – the potential to totally disrupt and transform business processes and established business models. But there are many obstacles to overcome, many of them conceptual and in stakeholders’ minds, so this technology will only come to near its potential over many years, possibly decades.

So for me, it’s clear that IoT is still largely at the articulation stage. The future is here already – in a way, it has already arrived – but it is not evident that organisations are widely ready enough to embrace and utilise it to anywhere near its full potential. From wearables and connected homes to cars, planes and trains broadcasting sensor data, a whole new world and computing paradigm is fast emerging in front of us, especially with data that will only continue to increase. So, if business executives want to make the most of this trend, then adjusting and realigning their strategies may become imperative.

Leave a Reply