A disruptive view on artificial intelligence in data analytics

In the era of constant development in which we live, concepts such as Artificial Intelligence and Data Analysis are increasingly present.

However, they are not recent - they were born with what is considered the pioneer of computer science, ALAN TURING, very important in creating the basis for the modern computer and responsible for decoding messages from the German naval fleet in the days of World War II.


One of its objectives was to test the ability of a machine to assume a behavior close to that of humans. Then, from there, patterns began to be searched for in the data, which is repeated, in order to decide which is the next best option.

This is, in general, one of the premises of Artificial Intelligence:

The way we managed to mimic what is our associative thinking pattern, allowing us to identify missing and complementary pieces by comparison with the various patterns, the path to follow, in a process called training algorithm

Therefore, based on what is known, it is possible to determine with increasing degree of certainty, what follows.

We give an example: when drawing in the air, a semicircle by gestures, our thinking leads us to realize that what is missing is the rest of the circumference.

Designing this logic in computing, with Artificial Intelligence, neural networks (which simulate the associative behavior of the brain), are trained to determine what the rest of the pattern is. It is important to realize that the accuracy is greater the greater the information they already have.

We will see the role that this ability can play in data analytics.

What can Artificial Intelligence do for Data?

With the computational capacity on the rise and the possibility of doing network training to determine what is happening faster and faster, we are already in a phase in which Artificial Intelligence is present in the daily tasks, in a transparent manner.

In fact, it is even surprising how machines can "think about the future" and help increase productivity.

Let's look at a case study: when we write an email, the following word was suggested. Today the suggestion is for a complete phrase / expression, starting with just one or two words. This happens based on the phrases normally composed.


Google Smart Compose example with words suggested by the sentence context.

Furthermore, it offers us the opportunity to reduce mistakes, that is, not to make the mistake made today tomorrow. As?

Through a published text, for example, Artificial Intelligence already allows, by grammatical analysis, to identify the feeling / sensation that is being transmitted, analyzing the response interactions.

Thus, there is an opportunity to change or improve strategies and methods.

But be careful: it is necessary to transmit feedback to the machine, which is good or bad, so that its ability to standardize is improved.

In practice, we are strongly moving towards a moment when Artificial Intelligence is already what we designate as a commodity - several sectors of activity already assume that data analysis solutions have Artificial Intelligence components in their modus operandi.

Differentiating approaches to the use of Artificial Intelligence

The process consolidated in the past to give research instructions to the machine, assume using a specialized language to explore the databases.

Today, there is a way for us to ask a particular question exactly as we searched it on Google and to understand the answer. This phenomenon is called Natural Language Processing, in English Natural Language Processing (NLP), which transforms information of a more technical format into a language understandable for human beings.


Imagine the following scenario: in a telecommunications company, we try to find out how many calls occurred in the last month between Lisbon and Madrid. In the past, waiting for an answer from someone who had specific skills in the data model could make the information no longer as relevant or necessary when it was made available.

Today, it is possible to use systems that quickly search the databases and answer the question asked in natural language: “Calls originating in Lisbon and destination Madrid last month”. The same question asked in a database language, would imply knowledge of the lexicon only usual for technical users:


Example of SQL database search - Structured Query Language.

Artificial Intelligence together with NLP already offer the user the possibility of being able to access in seconds a graphic answer to what they are looking for, and even narrow the search by filtering by other topics.

Because? The Artificial Intelligence engine has already “learned” that users who ask that type of question usually prefer the information presented in that specific form.

In addition, these solutions allow analysts to be able to explain possible variations of data more easily, discovering and joining the factors that originated them.

Data Management in the future

In addition to the importance and concern in ensuring that there is authorization to access the data, we tend more and more towards a society in which they have to present three fundamental characteristics:


  1. Be true and trustworthy - so that the decision is made without a doubt;
  2. Be available to consume - for each user according to security policies;
  3. Be updated and “fresh” - so that you can decide on recent evidence without the risk of being incomplete.


Fundamental characteristics of Information for Analytical Exploration

But there is more: allied to these characteristics, consumers expect DATA ANALYSIS to be facilitated, through more intuitive and simple to explore solutions, with more intelligence and the ability to generate insights that propose future research and also faster, so that information justifies decision making objectively.



About the Author

Luis Miguel Correia

Luis Miguel Correia

Unit Manager - Analytics and Digital Process Automation

With over 20 years of practice in IT, Luís played an active role in the management of teams and projects in 5 countries, based on a trajectory of learning, Management and E-business, with a consolidated experience in more than 25 projects.