Gerhard Svolba and I have been colleagues for over 20 years. He started at SAS in 1999, after working as a researcher and SAS user at the University of Vienna in the Department for Medical Statistics. Gerhard is a Data Scientist (and Analytic Solution Architect) at SAS, and also a part-time lecturer in data science at several universities. I caught up with him to talk about his career in data science, and his experience of teaching, including a guest appearance at his son’s primary school many years ago.
Gerhard, tell me about your teaching work. Where do you teach, and in what subjects?
I teach at three universities for applied sciences, in Styer (University of Applied Sciences Upper Austria), Burgenland (University of Applied Sciences Burgenland) and Krems (University of Applied Sciences Krems). I teach sales and marketing students at Krems and Styer, and process engineering students at Burgenland. Statistics is not their main subject, so I try to show them relevant analytical business case studies, and provide as much context as possible. I also teach at two classic universities, Johannes Kepler University in Linz and every second/third year at Medical University of Vienna. I teach the medical students in Vienna on methodology, and at Linz I teach statistics students a mixture of statistical programming and business cases.
How do you find balancing your work at SAS with your teaching?
I left the university to do more practical work. I love doing research and teaching, but I also had so many great ideas that I thought would be relevant for customers, and so I joined SAS. After a while, I realised that I wanted to document my work, and so I started writing books. Teaching and writing both give me a chance to share my knowledge with other people. I love seeing people get excited about using data science to solve business questions.
We both know how fast data science moves. Are you planning on writing any new books to cover new areas?
This may change, but I currently have no plans to write any books in the next three years. I don’t really have time with all my teaching. However, I am reusing a lot of the content from my books in articles and YouTube webinars, which has given me the chance to update it a bit (there are some examples here and here). That said, I think my books are still relevant. I’m not saying nothing has changed, because we are definitely talking less about data mining and statistics and more about machine learning and artificial intelligence. If I was writing my books now, I’d use slightly different language, but the content remains useful.
What are your current projects at SAS?
Recently, I have been worked with the International Institute for System Analysis (IISA), a worldwide research company, on a computer vision project to classify satellite pictures of the Amazon rainforest. The idea is to use images to see where there is a high probability of manual deforestation. We are using artificial intelligence methods like computer vision and training and identifying sections on the picture. Another interesting recent project was the COVID-19 portal for the German Ministry of Health to bring together data from different sources, and make them analysable at different levels. My job was to build a capacity simulation to answer questions about the number of people who might need admitting to hospital, how many would end up in intensive care, and how long they might stay, and to forecast the number of beds needed. I’m also doing some work with natural language processing area to support public organisations and authorities and automate certain processes around building services for citizens.
Based on your experience teaching the next generation, how do you see the future of the data science and AI?
Looking at my experience over the last 25 years, we are already able to solve certain analytical problems which were not possible in the past. However, when I teach my marketing students, I see the importance of teaching people about very basic concepts. It’s vital that people understand that even fairly simple analysis can provide new insights. You also need to be able to explain this knowledge to others. The second issue that matters is that as a data scientist, you don’t spend all your time building models. You also need to clean data, or talk to others about where to find the right data. Analytics is multidimensional, from data to business.
Gerhard, thank you.
