The Internet of Things + The BI Industry

We continue hearing about how the Internet of Things (IoT) will transform the way we live by connecting to every aspect of our days to the internet through new devices. Whether it’s a fitness tracker, smartwatch, a new refrigerator, or that fancy thermostat we just installed in our home, equipment manufacturers have been trying to seed the marketplace with connected devices.

Companies such as Google, GE, Bosch, Samsung and especially telecom carriers such as AT&T and Verizon are all focused on establishing their IoT business by focusing on devices, standardization, platforms and infrastructure to support these applications.

A NEW KIND OF CONVENIENCE

Take your smart refrigerator, for example. As a consumer, we can breathe a sigh of relief knowing that our fridge will automatically order a replacement water filter at the right time, giving us more time to binge watch Game of Thrones or simply have one less thing for us to worry about.

However, for the manufacturer, this use of a connected device is even more beneficial than to the consumer. This scenario assures that consumers will automatically purchase replacement filters directly from the manufacturer at a point in time determined by Samsung, GE or LG, when the same consumer may have previously purchased generic replacement filters on Amazon.

Don’t get me wrong – I’m not suggesting that a connected fridge bears no benefit, but the Internet of Things’ value really comes to life when we harness the power of predictive and prescriptive analytics.

#IoT value really comes to life when we harness the power of predictive & prescriptive analytics.
_{Click To Tweet}

Before we get there, there are a few challenges we need to address.

THERE’S A LOT OF DATA TO MANAGE

Connected devices generate data – a lot of data.

In a connected thermostat, for example, it’s not unreasonable to assume that it will continually measure and record attributes such as current temperature, humidity, settings, battery level and so on. All this data is sent back to the manufacturer and has to be stored somewhere. Furthermore, the data has to be processed before we can even start thinking about applying any sort of analytics.

When you consider millions of thermostats creating hundreds of data points per day, we can quickly understand the magnitude of this challenge.

Hadoop alongside cloud providers, such as Amazon Web Services and Microsoft Azure will play a big role in IoT with their ability to store, scale and process massive amounts of data.

Amazon, for example, is offering increasingly powerful services through their Internet of Things platform, including Kinesis (which is a real time streaming service), S3 (Simple Storage) as well as DynamoDB, which is a massively scalable NoSQL database.

IoT companies are currently investing heavily into infrastructure to support their endeavors and have some of their brightest minds dedicated to streamlining processes around persisting and processing their connected device data.

In the coming years, after the foundation is set for data storage and processing infrastructure, we will see this work come to fruition with more of the predictive and prescriptive analytics offerings.

PREDICTIVE ANALYTICS + IoT

Let’s put predictive and prescriptive analytics into context in an industrial setting.

If I’m the operations manager in a plant, my priority is making sure that uptime is at 100%, maintenance cost is kept to a minimum and my team has a safe working environment. In a plant where my assembly lines, motors and other moving parts are connected to the internet and performance data is sent back to the equipment manufacturer, the manufacturer can leverage analytics to predict, based on millions of data points from other similar machines, that a motor or piece of equipment may be failing soon.

In fact, based on specific device attributes and data about similar devices, an equipment manufacturer may even be able to pin point a precise future time of failure if the machine continues operating at current levels.

Knowing when something is about to happen to my machine or knowing when my machine actually needs maintenance will help better plan and avoid costly unplanned downtime or costly routine maintenance that may not be required at ‘routine’ levels usually specified.

PRESCRIPTIVE ANALYTICS + IoT

Prescriptive analytics goes one step further by recommending (or prescribing) an action to keep the machine from failing. This could mean scaling back the RPMs of the machine while informing the operator that a service call is necessary. If a catastrophic failure is predicted, the machine could be shut down prior to such failure to avoid higher replacement cost, injury or even death.

However, Prescriptive Analytics does not just rely on what has occurred in the past or is likely to happen in the future, but takes desired outcomes, specific scenarios as well as all current and historic data into consideration when making a recommendation. This will allow us to prescribe the optimal way to handle a future scenario and goes far beyond the example of a maintenance issue or machine failure.

In the case of a manufacturing plant, prescriptive analytics could prescribe solution to a machine needing to be shut down, such as a change in setup to have other machines pick up the slack. This can be done instantaneously based on a large number of inputs with the desired outcome of having the least possible effect on product quality and quantity produced.

THE POWER

The Internet of Things, alongside the power of predictive and prescriptive analytics, has the capability to innovate our lives not just at home but also at work and in industrial settings, saving time, money, injuries and lives by being able to predict events and prescribe solutions.

IoT along with predictive and prescriptive analytics can save time, money, injuries and lives.
_{Click To Tweet}

Furthermore, it will allow us to move away from complete reliance on scheduled maintenance and running machines until failure, to detecting anomalies and being able to predict events before they happen and prescribing work arounds, saving costly downtime, maintenance and lives.

WHY IT MATTERS FOR BI

Knowing that there are hundreds, if not thousands of articles on IoT out there already, you may wonder why I’ve decided to add to the pile. I recently completed a consulting opportunity with an IoT SaaS company and experienced first-hand the challenges, opportunities and impact of the Internet of Things.

As businesses continue to generate more and more data over time, chances are as Business Intelligence Consultants, we will continue working with larger and larger datasets, which will require new techniques to efficiently store, process and analyze the data.

It’s not like storing, processing and analyzing massive data hasn’t been done before. In fact, it is done every day by companies all around the world, including recognizable household names such as Netflix, Facebook and Google.

However, the impact of IoT and the general growth in data will require more of us to become familiar with building, maintaining and tuning environments which can handle this load efficiently. Technologies will continue emerging to spur this innovation and as BI Consultants, this should excite us and make us want to continue developing our knowledge and skillset.

BI experts must continually learn new ways to build/maintain environments to handle #bigdata. #IoT
_{Click To Tweet}

BI experts must continually learn new ways to build/maintain environments to handle #bigdata. #IoT

Click to Tweet

On a personal note, I truly believe that the Internet of Things has the capability to transform our lives, not just in the sense of making our lives less stressful by having appliances re-order their essentials, but we can also track grandma walking off with her cane and warn us if she hasn’t taken her medicine then automatically prescribe actions based on her condition.

We can make communities a better place to live by improving waste management, transportation and traffic. We can help farmers farm more efficiently while avoiding major disaster in industrial settings.

As BI Consultants, we are part of this evolution – let’s embrace it!