Data Analytics in IoT/M2M
IoT is 'the' happening thing right now and is expected to continue this way as we move more towards connected world. It was and is among the top spots in most of the Industry buzz word list published in 2014 & 2015. But not many, including the technology people or enterprises, are aware as to what IoT really means to them or to the economy, or how to monetize the immense volume of data generated by the constituents (namely sensors, devices, microchips) of IoT.
Let's look at some of the use cases in diverse industries where IoT can be deployed to get a perspective:
Global logistics operators who have a wide variety of fleets across rail, road, air and sea routes are now realizing that effective utilization of their fleets can save millions of dollars. For instance, volumes of data collected on speed, acceleration, braking, temperature, and fuel can be collected and analysed in real time to identify the areas where efficiencies can be improved.
Using IoT intelligence, logistics companies can see how the use of the right route and acceleration patterns can affect vehicle performance and fuel usage. In addition they can also discover the impact of driver performance and how his behaviour can not only affect fuel efficiency, but also the longevity of the asset in regards to maintenance. Also telematics, GPS data, and local map software can be combined to make companies aware in real time of routes that may be affected by traffic jams, speed traps, or weather conditions.
Real estate/Building management:
Imagine working in a skyscraper that adjusts temperature and humidity to suit the number of people in your office, provides access to designated places and keeps elevator and power outages to a minimum. Now imagine all this can be remotely controlled for all the buildings that are owned by an enterprise or run by building Management Company across the globe.
The concept of the Internet of things (IoT) where everyday things are connected to the Internet presents unprecedented opportunities for the management and operation of real estate. With IoT, any part of a building can become a point to capture and send data. This data when analysed and made actionable will have the opportunity to explore, relate to, and interact with buildings in amazing new ways, to move from building management to full building automation.
So what seems to be crucial in each of the use cases above? It's obvious, gathering and analysing data.
The IoT data that is gathered every fraction of a second can be complex. For ages enterprises have not completely exploited the vast amount of data that they gather on an on-going basis and now IoT will bombard with more heaps of data.
Of all the big numbers being thrown around about IoT - I picked the below
$15 trillion - the economic value expected to be generated by IoT by 2030
$5+ trillion - 30-40% of total IoT market which can be attributed to analytics
Types of analytics:
Data collected in IoT can be processed and analysed under two different methodologies called Predictive and Prescriptive analytics.
Predictive analytics: Predictive analytics utilizes a variety of statistical, modelling, data mining, and machine learning techniques to study recent and historical data, thereby allowing enterprises to make predictions about the future. The purpose of predictive analytics is NOT to tell what will happen in the future, its purpose is to just predict or suggest what might happen in future.
For ex: In the fleet management use case we mentioned above, predictive analytics can be used to suggest routes that could have traffic jams during certain period.
Prescriptive analytics: The emerging technology of prescriptive analytics goes beyond descriptive and predictive models by recommending one or more courses of action and also showing the likely outcome of each decision.
For ex: In the fleet management use case we mentioned above, prescriptive analytics can be used to recommend which routes can be avoided by the driver and which route can be taken based on the time of the day and also display estimated time to travel through the recommended route.
Decentralization of Analytics processing:
Real-Time Analytics done over large tracts of data that are streaming in from all connected devices like sensors, wi-fi connections spread across geographies will generate tremendous value with tremendous impact. Also it would be much more efficient to have a decentralization of data storage, processing and analytics since there may not be just not enough network bandwidth in the future to transfer all the data in real-time. For instance, think about a ship in the middle of the ocean - do you really want to transfer all of the (low-value) log data from all sensors, machines, switches, etc to a central data analytics installation? The costs of transferring data from all over the world in real-time to a central location is much higher than the savings through economy-of-scale of a centralized solution. Furthermore, network latencies and interruptions omit the usage of centralized solutions.
The Challenge for IoT Analytics vendors:
There is a challenge for IoT Business Intelligence/Analytics vendors to create new tools that not only allow companies to capitalize on their own data but also aggregate sensors data gathered from sensor networks, public and private clouds and provide embedded predictive and prescriptive analytics services to support the enterprise decision makers in crucial decision processes that reinforce their ability to continuously improve the company's financial performance, to keep the costs down, and increase customer experience.
To thrive in the new environment, enterprises need solutions that use in-memory computing to harness the power of Big Data and advanced analytics to help them draw insights from - and make them more responsive to - the needs of digitally connected customers.