Looking Back & Looking Forward
In this very first (official) blogpost of “The Outlier” I have the honour to introduce you to our first guest: Willem-Jan van den Heuvel. As one of the “founding fathers” of the Jheronimus Academy of Data Science (JADS) and Director of Education of the same educational institution, he shares his view on the current state of being at JADS. Further, he tells about the emergence of the exciting field of data science, what makes a data scientist excellent and where he thinks the industry is moving towards.
As part of the final interview round, the so-called lightning-round, one of the questions I asked him is: What is something you believe that very few agree with you on? Willem-Jan: “A comment I heard over and over again at the beginning of the JADS-founding process is: it is not going to work. Although rectors supported our vision, there was a consensus within the university community that setting up a new master program on an entirely new campus was not feasible. I really did not agree with them and thought: it is going to work. In the end, we have been able to realise our plan with a relatively small team.” Therefore, it’s no surprise that he is very proud of the final accreditation of the master program (only 3 months ago!) and the whole team who made this possible. “It’s really an achievement and far from trivial. We not only needed to prove market demand and the importance from a scientific point of view but also the need for a new campus in Den Bosch.” After all, the scientific quality often fell short in previous experiments with new university campuses. On the other hand, the fact that the percentage of IT companies in Den Bosch is highest of The Netherlands (with exception of the Randstad) clearly confirms why Den Bosch is a good location for our campus.
The History of Data Science
In 2002 Willem-Jan obtained his PhD in Computer Science at Tilburg University. Although an annual data related conference (“Very Large Data Bases”) has been organised since 1975, at the moment of his promotion nobody had ever heard of what we like to call “Data Science” today. “Nevertheless, the field of study has existed for a long time from both engineering (Computer Science) and math disciplines (Statistics & Econometrics). In recent years it has taken a flight and now several disciplines come together in the field of Data Science.” So I wondered whether he believes that Data Science is really something “new” or simply a rebranding of existing methodologies and tools. “Although I already learned about neural networks in my study, I think that Data Science is really a new research discipline. At the time analytics was less popular because batch processing took hours or even a whole night. Due to the increasing processing power you can actually apply these big data techniques now. This has really been a big turning point in my opinion. At the same time it’s more of an evolution than a revolution.”
“At the time analytics was less popular because batch processing took hours or even a whole night.”
In this dynamic, new and evolutionary field it’s important to stay up to date with the cutting-edge technologies and trends. He elaborates: “For example, I had to study large data processing frameworks such as Hadoop and Spark. I have done a lot of self-study by reading books (his book recommendation: Data Science for Business) and I regularly attended courses by companies such as IBM. So I literally went back to the classroom to learn more about these technologies. Staying up to date is very important, both for your education and research.”
Rigor & Relevance in Research
Next to his responsibilities as Director of Education, Willem-Jan is also a lecturer for the Data Engineering and Innovation Services graduate courses. In the past he taught a variety of courses related to Information Systems at the Erasmus University, Fontys Den Bosch, Tilburg University, Rotterdam Business School, TIAS and Antwerp Business School. What stands out in the this sequence is the focus on society and business of these institutions, while his area of expertise is more on the technical side. Willem-Jan explains: “In education and research I try to combine two aspects: scientific rigor (i.e. correctness, consistency and transparency) and relevance. Eventually, I want to teach subjects which have a certain relevance. For example, the deployment of big data technologies to solve very concrete problems in a scientifically responsible manner. Pure fundamental research without any relevance I don’t find so appealing and it gives me less satisfaction. That does not mean I do not appreciate it. For that reason, I am at the right place at JADS given the Data Science (technical) & Entrepreneurship (business/social) program. The same holds for most of my colleagues. At JADS we especially focus on industrial applications while fundamental research is typically conducted by both parent universities (TiU and TU/e).”
Another remarkable observation in his career path, is the combination of both management and research tasks (e.g. Managing Director European Research Institute in Services Science). He explains: “On the one hand, I really like building something from scratch. Having said that, if you look deep into my heart I have a strong desire to publish research proposals. Besides that, you are sometimes asked to fulfill management positions and you cannot always say no to that. But again, I work at a university because research and education is where my heart lies.”
Excellent Data Scientists
Willem-Jan has been closely involved in the development of the Data Science & Entrepreneurship curriculum which builds upon 4 pillars: Data Engineering, Math & Statistics, Business Analytics and the Socio-legal-economic Context.
Due to the popularity of data science and abundance of external libraries and packages, the use of advanced techniques has become accessible to a wide range of audience. You may ask yourself: to what extent is it really necessary to have a solid foundation in for example math and statistics? Willem-Jan explains: “You must be able to drive a car, but as a data scientist you should also know what goes on under the hood. This helps you understand how new technologies work in the future: is it the right approach, is it reliable and in which situation can I apply a specific algorithm? We hope to deliver students who are scientifically critical the moment they graduate but also maintain this attitude in the future.” To illustrate this he refers to a course on Hadoop he currently teaches: “In the end it’s nice if students know the trick (it has a fair market value), but I think it’s more important that they understand how it works methodologically and conceptually. That’s what they bring along and which they can benefit from in the long run.”
“You must be able to drive a car, but as a data scientist you should also know what goes on under the hood.”
In summary, at JADS academics primarily focus on solving practical problems, still a theoretical understanding remains essential. On top of this, I asked him what he thinks makes an excellent data scientist. Basically, he mentions 3 criteria:
- Analytical skills — mastering the exact sciences (statistical and mathematical models).
- Engineering skills— someone who is not only capable of analyzing (big) data but who can also turn data into a product or service.
- Domain knowledge — connecting your product or service with market demand.
The latter is the big secret according to him: “It’s relatively easy to come up with a new algorithm. A PhD researcher should be able to publish an article about an incremental improvement of an existing algorithm without too much effort. Building a product or service that is adopted by the European community and delivers social and economical value requires certain skills which not everybody possesses. Then you are excellent in my eyes.”
An Entrepreneurial Ecosystem
There are multiple excellent data scientists but what is quite unique to JADS is its entrepreneurial atmosphere. Even more, Willem-Jan says that the majority of JADS-professors has an entrepreneurial spirit: they have the ambition to be part of a data science start-up. As a good example, he refers to Germany where it’s very common that professors and students start a company together (e.g. at the Fraunhofer-Institut). The professors typically have an advising role within the company which is very fun to do according to his German colleagues. Although it has not happened at JADS yet, he is very open to replicating this model in The Netherlands and knows that many of his colleagues share this opinion.
Currently, he is not actively part of a company, though if he would start a company tomorrow it would be related to predicting and preventing subversion (e.g. weed cultivation, terrorism). So for those entrepreneurial “Intro To Data Science” students of you who picked the Global Terrorism Dataset and identified a golden nugget, you know who to approach now..
The Next Big Thing
As the falling paper price made it possible to collect and share more data in the 19th century (see 07:30 in his talk below), increasing computing resources now make it possible to execute advanced queries on large datasets as we saw before. What’s the next big thing?
“First, the upcoming trend will be the automatisation of data science techniques. I don’t see this as a threat, because it will only be possible for a subset of all tasks. It will make the work for data scientists more appealing since repetitive tasks will be automated while the truly challenging tasks will remain.”
Second, another industry trend which he thinks will continue is the implementation of more and more “Internet of Things” (IoT) devices to collect huge amounts of data from multiple data sources realtime. Here he does see a potential threat. For example, the cameras integrated in billboards at NS-train stations. “Then I sometimes think it’s going too far, especially because it comes so close to your personal atmosphere. As a society we should keep an eye out for this.”
Third, he truly believes Data Science is not a hype but is here to stay. Even more, governments and businesses will be more dependent on it over time. He expects more professions will require data science competencies. “As we have also seen in computer science disciplines, data science will be become an integral part of many curriculums such as healthcare and the police.”
“Data Science is not a hype but is here to stay.”
Finally, I asked him how he will use data science methods to improve education at JADS. By chance, he discussed the same question at the “Nederlandse Organisatie voor Wetenschappelijk Onderzoek” (NWO — Dutch Organisation for Scientific Research). There he talked about a better student tracking system, whereby data is combined from primary school to (applied) university. On the other hand, he sees potential to evaluate teachers’ performance and for example offer brush up courses wherever appropriate.
Hopefully, you found this article helpful in any way. If so, you could do us a huge favour by telling us what you liked about it in the comments down below. Also, recommendations for interviewees and column publications (yes, your very own story!) are much appreciated ❤️!
And of course I should not forget to thank my fellow interviewer for helping me conduct this interview. Thanks, Joseph 👏!