Back

 Industry News Details

 
Simplifying The Fourth Industrial Revolution: Data Science, IoE & AI Posted on : Mar 28 - 2017

We are in the midst of the fourth industrial revolution, a transformation revolving around intelligent machines. Some of the newer concepts are overwhelming by their sheer impact, and transformational technologies like Tesla and Siri are just the tip of the iceberg. As marketers, it is imperative that we understand these concepts and how they are impacting our lives.

What is data science?

One of the fanciest things about data science is the current salary being offered to the experts. But the real worth of data science comes from its ability to predict future decisions or incidents. It isn't magic, even if it seems like it. In actuality, it is a four-part sequential process, each one dependent on the previous part, that becomes more and more accurate with increased iterations:

  1. Data clarity: Data can come in the form of numbers, text, audio, images, video, or any combination of these mediums. It is the foundation of data science and the more clean and reliable it is, the better. We all witnessed presidential candidates in the U.S. fighting over each other’s data, and yes it is worth fighting for. If you don't have access to a reliable data source, you can't start on a project. Even if the data is complex, there is software that can put it into a usable format.
  2. Data analysis: Numbers are analyzed using the BODMAS rule, text is analyzed by reading, and richer media is analyzed through pixels. Whatever the data, analysis is the act of identifying patterns.
  3. Data interpretation: Interpretation first kicks in when we recognize relationships between patterns. It could involve repetition over time or purely a cause-effect relationship. A lot of us grew up making graphs, while highly-paid data scientists use advanced software to decipher complex equations to define a relationship.
  4. Data application: This is where real value and serious money is generated. The ability of a relationship to be extended into the future allows us to predict decisions. The ability to predict the future accurately generates money for an organization or individual, as Nate Silver did with Obama’s victory or as Google does when completing our sentences in keyword search.

The applications are limitless and opportunities to make money are burgeoning across domains. We will soon be able to hire the right people by accurately predicting cultural fit through talent science. We can predict the emotions that will appeal to our target audiences to position ourselves as emotionally appealing brands. The list is endless, as long as data science starts with reliable data.

Why is IoE such a big deal?

First there were computers, and they were connected to form the internet. Then more internet-enabled machines became connected. Now the number of devices connected to the internet is outstripping the number of humans. Many actions on the internet are initiated by humans, including content consumption, communication, commerce, social media, etc. However, there is also data being generated through inter-machine communication with no human interaction at all, such as remote switches, flight data, etc. This is referred to as the Internet of Things. Together, the two make up the Internet of Everything (IoE). This gives equal weight to data generated by machines and data generated by human action. Thus, human-machine collaboration has already begun.

There is, however, a big vacuum in data science on IoE. My company estimates that 90% or more of the data being generated is non-numeric, i.e., complex and unstructured. Traditional software companies have focused on writing code that helps businesses, and the IoE ecosystem is uncharted waters. This modern treasure hunt has many techno-adventurers innovating by analyzing and interpreting IoE data for predictive application. For example, I would love to see a Forbes interface that delivers multimedia content that is packaged identically for all subscribers but curated contextually for each individual subscriber.

What is artificial intelligence?

AI is the capability of machines to make decisions or act like humans. Jarvis from Iron Man is a good illustration of this concept. Machines are all programmed to be logical and identical. However, humans have different mindsets despite a somewhat uniform programming through education, faith, societal rules and experiences. This human context is captured by analyzing unstructured IoE data, which becomes a sort of contextual intelligence pool designed to simulate human actions. Thus contextual data science, using IoE sources, forms the backbone of modern day AI.

As I mentioned earlier, the big money is in the ability to predict actions and decisions accurately. And every action is now predictable. AI machines are self-correcting, so every error is analyzed and continuously enhances accuracy. A simple analogy is weather forecasting, which varies by region and time. But a combination of contextual understanding of the region and an analysis of error over the years has yielded very high accuracy weather forecasts at a city level. Similarly, digital marketers have been using contextual advertising through retargeting. AI takes this to a new level by doing contextual advertising right away, eliminating cumbersome cookies, which is both more efficient and effective.

In this autonomous era, there is a huge need for contextually analyzing IoE data and using intelligent machines to mimic human data science— and release huge business value.

If you have ideas on how best to contextually analyze unstructured data, please comment below. Source