The Data Scientist: a profession with future

Meanwhile, the profession has attracted public attention. But what is actually behind it? Dr. Benedikt Schmidt and Dr. Marco Gärtler from ABB describe the job and explain why you should be enthusiastic about it.

DECHEMA: ABB is partner of the KEEN platform, which focuses on the application of artificial intelligence methods in the process industry. What role do AI methods play in ABB's research projects?

• __ Dr. Marco Gärtler: In recent years, an incredible progress has been made in the field of artificial intelligence. For example, the AI researcher Geoffrey Hinton, professor at the University of Toronto, Canada, contributed to the breakthrough of artificial neural networks, which are used for deep learning. The technologies are now mature enough that industry sectors such as healthcare, manufacturing, mobility and construction are evaluating their usage, and are currently implementing them. As partner of the KEEN platform, ABB is developing various AI-based applications for the process industry.

• __ Dr. Benedikt Schmidt: For ABB, AI is considered as an additional tool for the method orchestra. They complement classic methods such as physical modelling (first principle model). Machine learning as a subset of AI is used to generate knowledge from data and to facilitate decisions based on this knowledge. This process always takes place in interaction with the other tools of the orchestra. Machine learning, for example, is able to identify patterns in existing data sets.

DECHEMA: As Data Scientist, both of you have already applied AI methods. What is behind the job description of a Data Scientist and what makes the profession so exciting?

• __ Dr. Marco Gärtler: The term “Data Scientist” has increasingly appeared in connection with big data, initially more in consumer applications and social media, but also in traditional industries. The core tasks of a Data Scientist are to analyse and evaluate data, gain insights, falsify hypotheses, and make predictions. Due to the higher availability of data in recent years, newer methods are also being used. The general availability of open source tools such as Google TensorFlow plays a crucial role for the success of the Data Scientist.
• __ Dr. Benedikt Schmidt: Female and male computer scientists, mathematicians, physicists, chemists, engineers, and neuroscientists work together at ABB's Corporate Research Center in Ladenburg. In the wide range of topics, which includes industrial robots, process automation, or areas such as electrification and sensor technology, new challenges are constantly arising. For example, one exciting issue for me in the field of predictive maintenance was to predict gearbox problems in robots. As a Data Scientist, you are constantly evolving and learning about new industrial applications, which is exactly what makes the job so exciting.

DECHEMA: What special qualities are necessary for the job?

• __ Dr. Benedikt Schmidt: Besides a technical understanding in STEM fields, communication skills are one of the most important soft skills of a Data Scientist. A typical project at ABB's Corporate Research Center can last from two months to three years. At the beginning of a project, the problem is often not fully specified and intensive discussions with the client about the problem, the data, the objective, and the domain are conducted. People with different backgrounds are involved in this process. Subsequently, data acquisition and an initial analysis are performed, where IT infrastructures and legal issues also play a role. Classical statistic is used to try to generate an initial understanding of the data, identify data structures and find representative data sets. At this point, the first data science methods are already applied. After the preparation phase, machine learning, for example, is used as a method to generate a model. Consultation with the customer or domain expert and the work in the team are essential here. In the following, the solutions are evaluated together with the customer or end user. In the final stages, a prototype is the output in ABB's Corporate Research Center. Scalability and IT infrastructures are addressed afterwards together with ABB's business units. Communication skills are also relevant here.

DECHEMA: ABB has formulated a question for the KEEN Hackathon. What is interesting about your challenge and why is participating at a hackathon a special experience?

• __ Dr. Marco Gärtler: Solving a challenge within a very short time in a team in the format of a hackathon goes beyond the professional and student experience. We also observe this in our ABB internal hackathons. Our challenge of the KEEN-Hackathon allows a wide variety of approaches for solving the problem and gives room to experiment in the field of industrial artificial intelligence. In daily practice, most collected datasets are "dirty" and need to be pre-processed manually. Successful machine learning in an industrial setting, however, requires cleaned and labelled datasets. The problem is rarely found in the public discourse. Our task of the automatization of data pre-processing is therefore highly relevant. For me, it was curiosity that drove me towards the participation in a hackathon for the first time. The profession of a Data Scientist has future. A hackathon can be a good introduction to the environment and incidentally offers the opportunity to get to know ABB as an employer.

The interview was conducted by Dr. Simone Rogg/DECHEMA e.V.