What are common mistakes when hiring data scientists?

To capitalize on mountains of data, data scientists are essential. Big data initiatives have so far resulted in an average increase in sales of 13 percent.

The information that is in their mountains of data gives companies, above all, information about their customers - their behavior, their preferences and routines. If you analyze the customer data, a huge economic potential opens up. Big data analytics has therefore developed into a key technology for almost every company.

The majority of company leaders are aware of the enormous influence that data analysis has on their own economic performance. According to the BARC study "Big Data Use Cases", data analytics lead to better strategic decisions for 69 percent of those surveyed. More than half stated that they could control operational processes better.

On average, the big data initiatives resulted in an increase in sales of 13 percent and cost savings of 16 percent. "There is no getting around big data", says BARC managing director Carsten Bange. "Our study clearly shows that big data analyzes give companies a competitive advantage and bring tangible, measurable benefits with them."

So the value of data analysis is now accepted - a central problem remains unsolved: Who should filter the corporate gold out of the data. So far, there is hardly any trained specialist staff available to analyze the data and convert it into business benefits. There is simply a lack of employees with the necessary specialist know-how to optimally exploit the big data potential. And across all sectors: banks and consulting firms are looking for analysts who specialize in large amounts of data, as are car manufacturers, insurance companies and administrative authorities.

These experts have been around for a long time - at least as a concept. Under the generic term data scientist - data scientist in German - a job description has been established in the IT world for several years that summarizes various role definitions and names from the big data environment.

In the zoo of the data specialists

activity content Required know-how
Data scientist Determines which forms of analysis are best suited and which raw data is required and evaluates them. Communicate the results Mathematics, statistics, databases, programming, business intelligence, psychology, media
Data artist Ensures the comprehensible presentation of the evaluations in the form of diagrams and graphics Graphics, statistics, presentation techniques
Data Architect Creates data models and determines when which analysis tools are used Databases, data analysis, business intelligence
Data engineer Takes care of the hardware and software, especially the analysis systems and network components Hardware, software knowledge, programming
Information broker Databases, communication, psychology
Data steward Addresses any questions and inconsistencies in the data sources. Is used in particular in data quality management To achieve compensation
Own research / Wikipedia
activity content Required know-how
Data scientist Determines which forms of analysis are best suited and which raw data is required and evaluates them. Communicate the results Mathematics, statistics, databases, programming, business intelligence, psychology, media
Data artist Ensures the comprehensible presentation of the evaluations in the form of diagrams and graphics Graphics, statistics, presentation techniques
Data Architect Creates data models and determines when which analysis tools are used Databases, data analysis, business intelligence
Data engineer Takes care of the hardware and software, especially the analysis systems and network components Hardware, software knowledge, programming
Information broker Databases, communication, psychology
Data steward Addresses any questions and inconsistencies in the data sources. Is used in particular in data quality management To achieve compensation

What do data scientists do?

The central task of a data scientist is to dig through vast amounts of data in order to gain business-relevant statistical insights. It defines which forms of analysis are best suited and which raw data are required for this. Using mathematical-statistical analysis methods, he develops models for information extraction and forecasting for big data applications. Ideally, the results flow directly into a tangible business plan.

The job requires profound technical knowledge in the areas of statistics, mathematics and probability theory as well as in programming and database applications. In addition, there are methods of machine learning, artificial intelligence and prognosis. The data scientist uses both structured and unstructured data as the data basis. He either carries out the data analysis himself or supports IT specialists with the analysis or guides them.

The evaluation of data is only one part of the field of activity of a data scientist. He also has to understand business contexts, needs industry knowledge, psychological knowledge and negotiating skills. And he should be able to clearly present the knowledge gained. Third parties should be able to understand its results without specific expertise. Because he acts as a mediator between specialist departments and management, good communication skills are also important for the success of his work.

Susanne Wolf, Chief Human Resources Officer at Alexander Thamm GmbH, one of the first German data consulting companies, sums up the role of the data scientist in the Unicum job exchange as follows: "A data scientist must have analytical understanding and be able to deal with the most common programming languages , Bring communication skills with you, appear smart and self-confident and be able to ask very specific questions. "And further: Data scientists often work in interdisciplinary teams, spend a lot of time in meetings, have to present their results in a visually appealing way and present them convincingly. The hard number crunching is only part of the job.

Data Scientist: "The sexiest job of the 21st Century"

The demanding all-round profile of mathematical-analytical, business, communicative and creative skills gives the data scientist an almost god-like status. "It is too early to describe the data scientists as the new masters of the universe. But they are on the way there," says Alexey Loganchuk, founder of the New York-based recruitment consultancy Upgrade Capital.

Loganchuk is not alone in his assessment. In its ranking of the "best jobs in America", the US rating portal Glassdoor put the data scientist at number one at the beginning of 2018 - for the third time in a row. The career network LinkedIn calculated a job growth of 650 percent for data scientists since 2012. And the renowned management magazine "Harvard Business Review" even named the work of the data scientist the "sexiest job of the 21st century".

No wonder that the demand for the new favorites in the IT scene is very high. According to management consultants at McKinsey, there were 150,000 open positions for data specialists in the USA last year. In Germany the demand is similarly strong. The Stifterverband für die Deutsche Wissenschaft estimates the need for specialists with data skills at up to 95,000 positions.

In the job advertisements of the Federal Employment Agency - not exactly the typical contact point for the placement of highly qualified IT personnel - 402 vacancies for data analysts were reported last year, including from hospitals, insurance companies, logistics companies and government authorities. At Deutsche Bahn alone, no fewer than 30 job postings were online at the same time in January 2018 - only for data scientists.

What do big data experts earn?

In large companies, there is particularly high demand for freelance big data experts. According to the Freelancer Index of the digital association Bitkom, four out of ten major German companies expect demand to continue to rise. From the companies with at least 2,000 employees even expect
45 percent an increasing demand.

The lack of data specialists is also noticeable in terms of salary. According to a study by US consultant Winter Wyman on around 620 IT positions, big data engineers earn up to US $ 15,000 more annually than user interface specialists and up to US $ 27,000 more than software engineers.

In Germany, the gap is not quite as wide. Frank Pörschmann, board member of the professional association of German data scientists, calculates that big data specialists in this country can expect up to 20 percent more salary than other comparable IT professions. However, large corporations would not pay significantly more for data experts than for other IT developers because of rigid salary schemes.

The freelancer placement portal Gulp provides current salary information for the freelance big data job market. After that, freelance big data specialists in Germany can expect an hourly wage of 90 euros. According to Gulp, this is how much IT freelancers currently charge for one hour of work if they specialize in big data. If all hourly rates for IT freelancers are added up, the result is 83 euros, which is seven euros lower.

The royal road: university studies

One of the main reasons for the lack of experts is the lack of training opportunities. Until recently, there was hardly any official degree in data science. The role was usually taken by lateral entrants from subjects such as business administration, computer science or mathematics. Many data analysts who call themselves data scientists today have acquired the necessary knowledge over the years through experience and initiative. But those times should be over.

More and more colleges and universities are now offering data science courses. It seems as if the demands of the economy for courses with a specific focus on big data have finally fallen on fertile ground. In the USA, the former US President Barack Obama declared data science to be the top priority in the education sector back in 2013. He pledged $ 37.8 million to universities for this reason to help advance education. With success: At universities like Stanford and UC Berkley, data science is one of the most popular subjects among applicants.

Here, too, those responsible for education have woken up. More than 20 universities and colleges in Germany and Austria already have data science courses in their programs. The forerunner was the Technical University of Dortmund, where the faculties of computer science, statistics and mathematics have jointly operated the data science course since 2002. Other universities have followed suit. The focus is on master’s courses, but you can also complete data science as a bachelor’s degree.

In addition to standard courses, some universities offer specialized advanced training courses and part-time courses - for example Albstadt-Sigmaringen University. The Stuttgart Media University also offers working people a master's degree in data science and business analytics - in addition to their professional activities.

Further education and training opportunities

However, studying is not the only path to a career in big data. In the meantime, a broad market for training and further education has established itself, with which IT and other specialists can expand their skills in a targeted manner. These training opportunities are supported by public and private research institutes, technology and software manufacturers, professional associations and other institutions.

In the higher education sector, for example, the Brandenburg University of Technology, together with the Agency for Scientific Further Education and Knowledge Transfer e. V. offers a "Data Science" certificate course. The Bitkom Academy and the Fraunhofer Institute for Intelligent Analysis and Information Systems also have a comprehensive seminar program for data scientists. The package is aimed primarily at practitioners and consists of numerous modules that can be combined depending on the focus of activity - project manager, developer, analyst, sales manager. Usually such training programs end with certification.

In addition, there are many further training opportunities from IT companies such as IBM, Oracle, HP or EMC. The EMC Academic Alliance, for example, offers a “Data Science and Big Data Analytics” curriculum. Online platforms such as Coursera, Udacity or edX also offer courses in which the basics of data science can be learned. Those who have the appropriate prerequisites can also acquire key qualifications as part of a trainee program.

These alternative further education and training opportunities are particularly interesting for companies that want to equip their employees with data science know-how. This is a good opportunity for engineers, economists, statisticians, mathematicians or similar experts who are already working in the company. They also have the advantage that they are firmly anchored in practice and already bring the relevant knowledge with them.

Find a data scientist

Because the labor market is currently finding it difficult to meet the growing demand for data science experts, medium-sized companies can hardly avoid training their own employees accordingly. According to Sopra Steria Consulting, a good half want to build up internal knowledge through suitable training measures for existing employees. The advanced training opportunities mentioned are a way of imparting the necessary know-how to employees.

If, on the other hand, you want or have to acquire big data experts on the open labor market, you should come up with some ideas.

Job advertisements should speak the language of the data experts and be precise in the technical requirements. It is not uncommon for experts to read from the job profiles that the potential employer himself has little competence - no data scientist wants to work there. In addition, specialized and innovative service providers can help find suitable data experts. And finally, good networking in the big data scene can also be important in order to come into contact with experts - or, conversely, to be courted as experts.

A possibly even more successful method of finding experts are data challenges. In these "competitions", companies provide young data talents with specific tasks to be solved via a cloud platform such as Topcoder or Kaggle. Developers interested in a job submit a solution and compete with the other participants. This type of expert recruitment has the advantage for companies that they can find dozens or hundreds of data scientists and developers who have been trained in the areas in question and have the necessary skills.

The automatic data scientist

Perhaps the scarcity problem will be solved by the very methods that data scientists have been using more and more recently: automation and artificial intelligence. Gartner estimates that by 2020 more than 40 percent of data science activities will be automated. The "Citizen Data Scientist" so dubbed by Gartner - in principle a

A layperson who is not primarily involved in statistics or analytics could then close the gap between the mainstream self-service analytics of business users and the advanced analytics techniques of data scientists. The narrow-gauge data analysts could carry out analyzes that a few years ago had a great deal of expertise

would have asked - but now without detailed specialist knowledge. "Most organizations do not have enough data scientists permanently available," says Joao Tapadinhas, research director at Gartner. "Instead, there are many skilled analysts who can become citizen data scientists. Equipped with the right tools, they can perform complex diagnostic analyzes and create models based on predictive or prescriptive analytics."

Frank Pörschmann is not quite so optimistic about Citizen Data Scientists. "Data analysis with automated tools by - more or less - laypeople only works to a certain extent," says Pörschmann. "With feature engineering, for example, this is possible. Here I can throw the data record into the machine and I automatically receive suggestions at the push of a button. The data must, however, be cleaned up beforehand - and the results are only suggestions. That certainly saves time. But you can Don't automate all of the work of a data scientist, and don't misunderstand the Gartner statement, because automation makes data analysts more efficient and teams more productive.However, it does not solve the growing need for data professionals. "

Degree programs for data science