According to Wikipedia, data science is an inter-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from many structural and unstructured data. All or nothing at all. This sounds like a chorus of a teen-pop song. The problem is that when the companies hire a data scientist, especially during the digital transformation phase, they what it all. What they really get instead is a statistician, a computer scientist, an analyst, a data engineer, an IT project manager, you name it. From my perspective we still miss the clarity to distinguish between operative data science and research and development driven data science.
An operative data science has a superior aim to support operations of a company. This person enables and ensures data-driven decision making on the operative level. And then there is a person who does all the fancy stuff, like machine learning and artificial intelligence – the research and development data scientist. Both may have similar training and experience but clearly different stakeholders. An operative stakeholder wants support to create and easily access good quality data for the reports and decision making. Nearly 90% of the time is spend by most data scientist in industrial companies to coop with data. Therefore 90% of your time, you are a data engineer. The reports and visualizations is the other part. These low hanging fruits are necessary, operations require them. Sometimes method equals business value. But more often the value is determined by the organizational needs of a company, not by the method. Dashboards and reports are relatively simple and easy from a data scientist perspective, but their impact is high.
A data scientist located in research and development department has stakeholders who would like to dig deeper, to support the vision of a company with strategic ideas, inspired and initiated by data. This may be a fresh data-based view on the customer structure, product portfolio or even production or distribution site. Assumed you have a good quality data, data science team can apply a huge variety of more sophisticated, machine learning and AI methods. Of course, these methods can be applied on the operative level as well. But the experience shows that the operative data scientist just does not have enough the time, from the development to the implementation of such products.
What is a data scientist? The answer is, it depends. But it is definitely not all in. We focus our training and experience on certain areas. Hence, dear companies, when you open a data scientist position be precise where do you want to locate a data scientist. If you are looking for an operational support, this person should be hands-on in data engineering techniques, analysis method and visualization. In research and development, you would require somebody who strongly advances in machine learning and AI.
Credit: Data Science Central By: Dr. Katharina Glass
