2 Types of Data Scientists and What Companies Need Them For

One of the key profiles in the companies’ templates and one of those that has become crucial is the data scientist.

Share Give it a Spin!
Follow by Email

One of the key profiles in the companies’ templates and one of those that has become crucial is the data scientist. This profile appeared in the heat of the big data boom, as an outstanding tool to optimize results and, above all, make the data serve for something.

The data scientist is the one in charge of profiling the information, reading it and extracting what the company needs from it. He is the one who clears the data and, above all, it is a profile that goes beyond the technical. In the first descriptions of these professionals, it was repeated that they were a profile “half artist, half analyst”. The training and experience requirements were from the beginning also varied and complicated, because not only had to have tech capabilities but also knowledge and talents in other areas.

But the profile of the data scientist is more complex than simply staying with the generality of what he does and what is expected of him. In fact, and given that in recent years they have had to assume more responsibilities and more work elements, one can already separate between two kinds of data scientists.

As they explain in an analysis in the Harvard Business Review, putting all the data scientists in the same bag would be too reductionist and a simplification of their work. It could be read, therefore, as a risk also for the decisions of the companies and for the work that these professionals do in them, impacting on their line of business and their relationship with consumers.

The two types of data scientists

The first large group of data scientists is modeling scientists, modeling scientists, who are responsible for doing, as they explain in the analysis, data science for machines. They are those who start from the data of consumers and the information they generate and design algorithms, models or training data. These are the professionals who create recommendation systems in online stores, for example, or those who create patterns and models that allow more clicks online. His profile is eminently technical and his work is in the area of ​​technology.

The second great group is that of the decision scientists. They are the ones who do data science for humans and their work puts them on the same level as the designers, the product managers or the executives of the companies. What they develop and what they conclude has direct applications on people.

These are the experts who use the data that consumers generate to understand them better and make decisions about what to sell, how to do it or when to do it. For example, it is those who analyze which action achieves better sales data or in what terrain of the customer experience we must work to improve it in general.

The work of both is similar, in the sense that both work to improve the position of the company, but does not go on the same side or have an impact on the same elements or in the same areas. Therefore, companies have to think very well what kind of professional they need and in what area they need to act.

One more complicating element

Knowing what data scientist is what is needed and what is sought helps to position the company much better, since this professional will work on the exact ground that needs to improve, change or act. But, at the same time, adding this layer of specialization, however inevitable, also makes things more complicated. If it is already difficult to capture talent based on a general premise, more is based on a specialization of specialization.

And, in addition, as companies become larger and the role of their data scientists more decisive, they also appear what the analysis calls ‘gray areas’, spaces that are neither one nor the other type of data scientists, but they are still necessary so that they can do their work. This is what happens with those responsible for data infrastructure, quality, management or those responsible for data engineering, for example.