Data Science and Machine Learning [#3] – Business Intelligence, Machine Learning, and Artificial Intelligence

In this post, I will show you how to add business intelligence (BI), machine learning (ML) and artificial intelligence (AI)  to the diagram (image above) that we built in previous posts. First, let’s focus on what is business intelligence and how does it fit into the diagram.

Business intelligence is the process of analyzing and reporting historical business data after reports and dashboards have been prepared. In other words, BI aims to explain past events using business data. With this information, it’s possible to figure out where the BI rectangle should go. It must go on the left side of the timeline, as it deals only with past events, and it must be within the data science rectangle (green rectangle).

As a subfield, BI fits comfortably within data science because it’s the preliminary step of predictive analytics.

To create appropriate models that could predict the future of your business accurately you must analyze past data and extract useful insights using these inferences.

Now, let’s see how we should distribute the three terms that remain on the left side of the vertical line (preliminary data report, reporting with visuals and creating dashboards).

The preliminary data report is the first step of any data analysis and it can also be considered as data science, so this term can stay where it is. As with reporting and creating dashboards, these are precisely what BI is all about. So we will place these two into the brown rectangle.

Now we delve into the controversial yet rapidly expanding field of AI and its subfield ML (the ability of machines to predict outcomes without being explicitly programmed to do so). This is about creating and implementing algorithms that let machines receive data and use this data to make predictions, analyze patterns and give recommendations on their own. ML cannot be implemented without data, hence it should stay within data analytics (blue rectangle) completely.

Some argue that data analytics and ML are two unrelated scientific fields for the sake of argument. I will let the ML and data analytics rectangles overlap and in a moment I will exemplify some areas that could be considered parts of both disciplines.

Moreover, ML should expand slightly to the left of the vertical line. The reason for that is the increasing tendency towards applying ML tools to the context of BI.

By definition, AI is about simulating human knowledge and decision making with computers. It is quite a general term that can have a rather philosophical interpretation.

We as humans have only managed to reach AI through ML, the discipline we just talked about and as the data scientists, we are interested in how tools from ML can help us improve the accuracy of our estimations. Let’s provide a few examples.

The demand for accurate real-time dashboards opens space for ML applications. ML software can hold data from third-party companies such as Facebook to detect new patterns from their data and thus suggest real-time recommendations and insights to managers and other decision-makers. This is a great way to improve the performance of a business.

Turning to the other side of the vertical line we can say that client retention and acquisition are two typical business activities in which ML is involved. It helps to develop models that predict what a customer’s next purchase will be.

As another example, ML can be applied to fraud prevention. We can feed an ML algorithm with prior fraudulent activity data and it will find patterns that the human brain is incapable of seeing. Having a model that can detect these transactions or operations in real-time helps the financial system to avoid a huge amount of fraudulent activity.

These examples are probably not the first that would spring to mind when talking about AI and ML. Usually, speech and image recognition are among the most popular examples as they are already being implemented in products like Siri and Google’s assistant.

Naturally, these two examples are outside the business context and it is debatable whether they could be placed under the data science, data analytics, both or neither as it is a hot topic for discussion right now.

I think they are worth mentioning, however, to avoid further confusion in disputes let’s take speech and image recognition out of the picture.

Finally, an example that is considered AI but not ML is symbolic reasoning.

It is based on the high-level human-readable representations of problems in logic. It was once a trend in the past when people were trying to create human-like intelligence. Today through ML is the only form of general AI that is being applied and symbolic. AI is rarely encountered let alone practiced. So, along with speech and image recognition, we can remove symbolic AI from the equation.

Remember that all areas intertwine in what we show here is not a strict representation of commonly accepted meanings and definitions. It is all a matter of interpretation and this diagram is a possible vision of data science. The locations of some of the components in the diagram can be somewhat controversial.

Bibliography

[1] Udemy – The Data Science Course 2020: Complete Data Science Bootcamp – 365 Careers
[2] https://www.toptal.com/insights/innovation/ai-vs-bi-differences-and-synergies
[3] https://expertsystem.com/machine-learning-definition/
[4] https://www.klipfolio.com/resources/articles/what-is-business-intelligence
[5] https://pathmind.com/wiki/symbolic-reasoning

Please follow and like us:
Bruno Silva

Author: Bruno Silva

Portuguese Electrical Engineer with special interest in Data Science, Robotics, Automation and Embedded Systems. With a Master’s degree in Electrical and Computer Engineering in the field of Energy and Automation.

1 thought on “Data Science and Machine Learning [#3] – Business Intelligence, Machine Learning, and Artificial Intelligence

Leave a Reply

Your email address will not be published. Required fields are marked *