KDnuggets Home » News » 2016 » May » Tutorials, Overviews » What are the Challenges of the Analytics of Things? ( 16:n19 )

What are the Challenges of the Analytics of Things?

without the AoT, it is difficult to realize the full potential of the IoT. We review the promise and challenges of Analytics of Things, including data, security, analytics implementation, standartization, and more.


After the Internet of Things (IoT), the Analytics of Things (AoT) is the next logical step for the enterprises. In fact, without the AoT, it is difficult to realize the full potential of the IoT. It is not enough to just accumulate a lot of data from devices, but enterprises need to make sense out of the data and do something that makes these devices more efficient. Also, the data generated have the potential to improve a lot of things about the business. This is where AoT is so relevant. Businesses need sound analytics that improves the ways it runs the business and overall, the bottom line.

However, as enterprises plan to turn to AoT, they face numerous challenges on the way. IoT itself is still evolving and AoT is in its infancy, so there will be a lot of confusions and misconceptions leading to wrong investments of money and effort. Enterprises need to invest on technology and skilled manpower to get the best out of AoT. A lot of time and patience is required on the way. The question will be, how many can sustain the tempo for that long?


AOT – What does it actually mean?

Analytics of Things is nothing but IoT analytics. In simple terms, AoT means generating analytics from the data generated by the IoT. IoT means that several devices are connected to the Internet and are transmitting data to somewhere. Now, just obtaining the data is the first step. Enterprises need to analyze the data to make the devices smarter and more efficient. The result of IoT analytics is also used to make right decisions in different situations.

Now, if we exclude the ‘Things’ part from the term AoT, then the rest is only ‘Analytics’ ,which is quite similar in nature to any other data analytics. Here the ‘Things’ are nothing but IoT devices.

Similar to other data analytics, AoT can be of different types like descriptive, diagnostic, predictive or prescriptive.For example, diagnostic and prescriptive analytics can be done with the help of medical IoT devices and predictions can be made based on the data generated by the industrial IoT devices etc. But, we must remember that all these forms of IoT analytics/AOT are still evolving and requires significant amount of time and effort to get real business value.

What are the challenges?

When we talk about the 'Analytics of Things', there are mainly two parts in it, one is the analytics part and the other is the data collection part, generated by the things/connected devices. The analytics part is reasonably matured but the biggest hurdle is the data collection part, which the analytics world is facing for years. So, we are actually iterating the same old problem while pursuing AoT. Analytics people might have a lot of innovative ideas about analyzing the data and getting wonderful insights from it. But the ground reality is, unless we have a proper infrastructure and skill to acquire and analyze necessary data, AoT is meaningless.

Now, let us divide the challenges into two broad categories, one is on the organizational side and the other is on the technology and implementation side.

Let’s start with the organizational challenges first.

The most important challenge is to build a solid AoT business case to convince the organization. It will ease the investment and future nurturing of AoT vision.The first investment is required to deploy the IoT devices in proper places with sensors to capture data. Once the devices are ready, organizations need to enable the data movement from sources (IoT devices) to destination (may be a staging DB or data warehouse or some other storage). Finally, a proper strategy has to be built to figure out how the storage and analytics can be managed.

Now let’s talk about some of the challenges on the technology and implementation side.

  • Data challenge: The volume of data each sensor generates is huge. But, the question is, are all these data worth transmitting?  The answer is 'No', so we need to figure out, how intelligently we can transmit only necessary and meaningful data. It will result a clean analytics without processing junk data.
  • Security challenge:Security and privacy of sensor generated data is very important. Specially, when this data is generated from sensitive devices fitted in confidential or critical areas. For example, the data may be coming from some devices fitted in an ICU (Intensive Care Unit) or from an airport or it can be from some critical industrial infrastructure. In all these cases, data security has to be ensured to protect the integrity of the system.
  • Analytics challenge:It is more related to filtering the entire analytics process. The challenge is - where we can do all these analytics? Maybe, some part can be performed within the devices, so that the data coming out of these devices are filtered to some extent. Or, we can design separate analytics layers once the data is reached unchanged, and then perform the filtrations step by step. And finally, do the analytics with the clean data.
  • Standardization/protocol challenge: Standardization and protocol is one of the biggest challenges for AoT success. We need to standardize the communication protocol between devices. It will help all the devices to communicate with each other seamlessly.

Apart from the above issues, we are also going to face a lot of new challenges in the coming days. As we move forward, we will have new IoT devices, new data format, new protocols and many more. So, it will eventually bring new hurdles to overcome.

What are the promises?

Like any new vision, AoT also has a lot of promises to fulfill. Although, the worth of AoT can only be realized with time, but we already have some examples in place which are really promising. The predictive analytics of AoT has proven its worth in places like ATM machines, or network systems. Self-driving cars, traffic information system are some of the other areas where AoT is already in place. It is also entering into medical industry, oil industry, fitness sector etc.

The AoT will grow as IoT grows. According to Gartner prediction6.4 Billion Connected "Things" Will Be in Use in 2016, Up 30 Percent From 2015’. So, AoT is believed to have a lot of potential and it will soon become a part of our lives.

The way forward

From the above discussion, we can easily understand that in the complex environment of IoT, AoT is a new entrant and just started to evolve.So, the obvious challenges of implementation will be there to overcome. Some of the challenges are on the business side, which can be managed with a proper use case and justification. And, the technological challenges can be solved with proper strategy, technology platform and skilled resources.

Organizations need to realize the potential of AoT and put proper IoT infrastructure in place. If we look back, then we can easily understand that the early adopters of big data have gained substantially. They were able to take the competitive advantages and gain in business. The same is true for IoT followed by AoT.AoT is going to be in the mainstream in next couple of years. So, it’s the right time to take the plunge and make a successful AoT vision in place, otherwise it may be too late to join the race.