What does a senior data analyst do

Data Scientist: What does the professional data evaluator do?

According to a LinkedIn study, the data scientist is one of the top 15 jobs for 2020. With 8th place, the professional data evaluator is well ahead. Here you can find out more about the job description and why young graduates are increasingly choosing a job in the field of business intelligence.

Data scientists not only analyze data, they also make forecasts based on modeled data.

Photo: panthermedia.net/Kheng Ho Toh

Data is worth gold. But only if you evaluate them correctly. This is where the data scientist comes in. He not only has to be familiar with statistics, but also with the matter he is investigating. Otherwise it can quickly lead to spurious correlations.

When traders want to get rid of their goods towards the end of the season, they start reducing them. "In technical jargon, one speaks of transcripts," explains David Krings, data scientist and co-founder of the AI ​​platform Aifora GmbH. Certain rules apply: An article once written off is never increased in price, but only further reduced if it is. For this reason, it is very important to choose the right time and the right amount of reduction. “If I reduce too much too early, I may have sold the product before the end of the season and not at the best possible price. But if I reduce too late and too weakly, I may not be able to get rid of the goods, ”says Krings.

What does a data scientist do?

The task of the data scientist is now to develop models that calculate the perfect time and the right amount of reduction. “In order to be able to identify a certain regularity, data from at least the past three years should be used,” explains Krings. However, the historical data must first be cleaned up and converted into the appropriate format. In particular, this means that data has to be mapped to one another and corrected from incorrect and inconsistent data.

On the basis of this information, the data scientist develops a mathematical model that contains and weights the relevant influencing factors for purchasing behavior. These models can be used to carry out simulations on historical data that show the potential of intelligent and automated markdown control. “It is important that the models are validated through tests and training so that the result is meaningful and corresponds to reality,” explains data scientist Krings.

Programming skills are a must for the data scientist

A data scientist should have a solid understanding of statistics and stochastics as well as programming skills. Important programming languages ​​are in this context

  • "R"
  • python
  • Matlab
  • Julia
  • Scala

The first two are fundamental, the others can be added depending on the industry. Krings has a bachelor's degree in “Scientific Programming” and a master's degree in “Technomathematics” (the course is now called “Applied Mathematics”). Krings also worked as a student assistant at the Supercomputing Center Jülich. There he dealt with programming on high-performance computers. These are computers that are able to process extremely large amounts of data.

The job was a good preparation for his later occupation. Because even a data scientist must be able to handle very large amounts of data.

Data Analyst: Preventing spurious correlation with engineering knowledge

“There are now applications that suggest algorithms so that everything no longer has to be programmed by hand. Nonetheless, a data scientist must of course understand how these models and the underlying algorithms work, ”says Tom Becker, managing director of Alteryx, which specializes in data science and analytics products. This is where the job description of the data scientist overlaps that of the data analyst. The latter also looks at data and scours data volumes for trends and relationships. But only the latter programs and models the data into a statement.

For Becker, engineers are the ideal candidates for the profession of data scientist. “Engineers are good at dealing with numbers. At the same time, they have a wide range of specialist knowledge, for example in automotive engineering ”. This enables them to assess which data are necessary for the analysis and which sensors have to be installed in order to collect this data. In addition, her specialist knowledge enables her to distinguish sense from nonsense.

Because it can happen again and again that facts that have nothing to do with each other develop coincidentally parallel to each other and are therefore brought into a wrong context. The best-known example of such a spurious correlation is the relationship between the stork population and the birth rate. In fact, the birth rate is higher in the country than in the city. There are also more storks in rural areas, but this does not necessarily mean that the birds bring the babies. A data scientist must therefore be able to interpret the statistical models or to build them in such a way that such nonsense does not even come out.

How do you become a data scientist?

In summary, one can say: Masters and MOOCs make the data master. However, there are two basic ways to work towards becoming a data scientist. The classic way is via a University degree. There are over 25 specialized courses and other combination offers across Germany. The following universities offer data science as an undergraduate degree with a Bachelor of Science degree:

Subject areaUniversityTime model
Applied data scienceUniversity of GöttingenPart Time Full Time
Data scienceHS baskingFull time
Data scienceHS Otwestfalen-Lippe in Lemgodual, full-time
Data scienceUniversity of MarburgFull time
Data scienceHAW XU Exponential in PotsdamFull time
Data scienceAKAD University of Applied Sciences Stuttgartextra occupational, distance learning, part-time
Data scienceUniversity of StuttgartFull time
Data scienceWest Saxon University of ZwickauPart Time Full Time
Data Science and Artificial IntelligenceUniversity of SaarlandPart Time Full Time
Data science in medicineTU UlmFull time
Digital Business & Data ScienceHAW Europe in BerlinFull time
Digital Business & Data ScienceHAW Europe in HamburgFull time
Mathematical Data ScienceUniversity of GöttingenPart Time Full Time
Medical Engineering and Data Science TH Aschaffenburgfull time
Business informatics - data scienceDual HS Baden-Württembergdual, full-time

Source: University Compass of the German Rectors' Conference

In addition, engineers can rely on what are known as Massive open online courses Further training at Udacity, the Decar Academy or the Unit-Me platform. "In the meantime, many German companies, especially in the automotive sector, are already working with such platforms," ​​explains Becker. He puts the effort to train an expert internally by means of training courses at around 600 hours.

“There is also free content on these platforms. These offer a good opportunity to test whether the professional environment could in principle suit you, ”advises Becker.

But it doesn't always have to be a straightforward path. Those who first opt ​​for a classic engineering degree can still follow the data path. The exemplary study and career path of Niels Röschmann shows how this works.

Niels Röschmann, 26 years old, now works as a digitalization engineer at EnBW Energie Baden-Württemberg AG in Hamburg. His career path there was anything but straightforward. After a voluntary ecological year, he began to study meteorology at the University of Hamburg. But this course could not inspire him at all. He moved to the Hamburg University of Applied Sciences and completed a bachelor's degree in the field of renewable energies. “I had very little computer science in this course, mainly electrical engineering and energy, of course,” says Röschmann.

Why the 26-year-old does not work as a classic electrical engineer today, but rather juggles with data, came about through practical experience. Röschmann started an internship with his current employer EnBW in the offshore wind energy sector. So he was able to observe directly how electrical engineers act in everyday life. He quickly realized that he was more interested in working in data management.

“I found it particularly exciting how you can advance the topic of sustainability with data analysis,” explains the student trainee at the time.

Now the question arose: How do I develop into this professional field? The universities seemed rather demotivating at this point. When applying for a master’s degree, the young engineering candidate found that he was not given enough specialist knowledge. "So I tried it at the Wedel University of Applied Sciences and actually got a place in the master's degree in IT engineering," reports Röschmann.

When asked why he finds the data scientist job more exciting than a job as an electrical engineer, he says that, in his opinion, the technology is already so mature that there aren't too many challenges waiting for you.

“The evaluation of data from the offshore wind farms gives me new tasks every day. Processes can be digitized or I analyze how the wind farm can run even more efficiently. That excites me. "

How much does a data scientist earn?

The starting salary of a data scientist is between 44,000 euros and 68,000 euros gross per year. A junior data scientist earns an average annual salary of 56,000 euros. The earnings are thus around 4,666 euros per month. Both Stepstone and Alphajump come to the conclusion.

Experienced data scientists can earn a significantly higher annual salary. This depends primarily on the training, the location, the industry and the content. If the data scientist is not only familiar with big data, but also with the product to be analyzed, he immediately has a better negotiating position.

More e-learning courses for engineers

There are jobs for data scientists in our job exchange

A contribution by:

  • Sabine Philipp

    Sabine Philipp has been working as a freelance journalist since 2004. Her focus is on technology, industry and business. In her article she likes to deal with the practical implementation of innovative technologies and legal requirements.

  • Sarah Janczura

    Sarah Janczura is content manager and responsible editor for ingenieur.de. After an internship with a focus on social media, she worked as an online editor in a digital agency. She writes about technology, research and career topics.