The need for data scientists has risen sharply in recent years. Reason enough to take a closer look at the tasks of a data scientist.
IT-related work areas play a central role in the success of companies today. The job profiles are just as diverse as the tools and programs that are necessary for performing individual IT jobs.
One area that has steadily gained importance in recent years is data science. With the modern possibilities for data processing, the need for data scientists who analyze both big and small data and make the results usable for companies has increased by leaps and bounds. That makes the job attractive for many. The two Harvard Business Review authors Thomas H. Davenport and DJ Patil even described the profession in 2012 as the “sexiest job of the 21st century”.
Put simply, the job of a data scientist is to extract business-relevant statistical insights from data. They assess which analyzes are best suited to generate insights and which raw data are required. Using scientific data analysis methods, they develop models for information processing and forecasts for big data applications.
However, a data scientist has to be able to do even more and, as an all-round talent, master various areas of work. Data scientists need to find out what the data is really telling a company, and they need to be able to differentiate between relevant and irrelevant data.
In an interview with ComputerWeekly.de, data scientist Adriana Menegozzo explains what tasks a data scientist has and what skills are required. Menegozzo was born in Italy and moved to Germany after studying mathematics. For the past five years, she has worked at Data Insights GmbH as a data scientist consultant and has supported customers from the idea to the implementation of machine learning and data science projects.
What are the tasks of a data scientist?
Adriana Menegozzo: The data scientist is a fairly new role that is still trying to find its right meaning. But I admit it is quite difficult to grasp the full spectrum of a data scientist’s job. The role varies a lot and requires a lot of flexibility. In general, you work with data, i.e. our present and our future, and try to gain insights and values from it. Ideally, you develop an algorithm that predicts what will happen in the near or distant future, recognizes patterns, and supports people in tasks that a computer can do better.
So basically you are the person who finds insights into the data, tries and explains algorithms, and empowers users to use the tools you have created. You have to be curious, read a lot, understand what’s going on around you, have good software development skills, and build as much knowledge of statistics, math, and machine learning as possible.
In which areas does a data scientist work?
Menegozzo: The great thing about this job is that it is very diverse. I mean, think of an area where knowing what is going to happen in the future is not critical. You can work in manufacturing, marketing, business, politics, health, communications, IT or retail. You will find data scientists in every sector. But most importantly: never work in isolation! If you want to do great data science work, you need a creative, diverse and competent team.
Which skills and characteristics or which training should a data scientist have?
Menegozzo: As the field of data science is growing incredibly fast, there are a lot of specific university degrees popping up, for example specializations in data science or machine learning, so I would say this will be the most common path going in the future. On the other hand, there are the key skills of communication, presentation, teamwork, willingness to learn and the ability to understand what is going on around you.
I think it’s an interesting field that opens up a lot of opportunities, especially for women. Since the data scientist is a new role, there is no gender bias like there is with other technical roles. This is an excellent opportunity to create a more diverse and inclusive environment in which women can also become role models. I hope and believe that we will be able to play a huge role in data science and we are the difference that can do it better.
Which companies employ data scientists?
Menegozzo: Of course, data insights. Seriously no: I think one of the best ways to find interesting companies and projects is with boutique consulting firms that are agile and not too big. Data science is a broad field with varying degrees of specialization: If you want to get really technical and develop the algorithms of the future, then start-ups or special research and development departments are perfect.
Large and medium-sized companies are generally looking for data scientists to improve their processes or their product. So for a more business-related position, this is the way to go. Ultimately, however, there is no such thing as a perfect company for a data scientist. You should find the one that is perfect for you.
Which tools and programs does a data scientist use?
Menegozzo: It depends on the position or the company. But I can tell you what I am using. I spend most of my time programming: Databricks, Spark and Scala when working with large amounts of data, Python and Bash for smaller data sets.
In general, you should familiarize yourself with machine learning algorithms and libraries, for example SciPy, Scikit-learn, Tensorflow, PyTorch, Keras, as well as with visualization tools. Remember, data engineers and data professionals are your best friends, so try to speak the same language and know the tools they are using.