The Future of JADS
Last October I got the opportunity to sit and chat with Arjan van den Born, the academic director of the Jheronimus Academy of Data Science (JADS). What follows are the highlights of our conversation. It will cover the fundamental insights he learned from engaging with data, the sciences and entrepreneurship. From humble beginnings, a fresh grad clueless of what’s next, to a family man following his passion in academia and entrepreneurship.
0. Academia to Expertise
Who is Arjan van Born? Well, the first thing you do is google your subject to see what the web thinks. It’s a short review of the top links, a few obscure ones on page 200 of the search results and an objective investigation of their social media (or you could just build a web-crawler to do the dirty work too). To summarize, and save you some time: he studied at various prestigious universities where he received a Bachelor of Arts, a Master of Science in Quantitative Economics and ultimately a PhD in Business Administration. I’ll leave the exhaustive details for you to read on his LinkedIn page.
Research and Expertise
When speaking to Arjan, I noticed his idea of a side-projects is writing academic papers, despite that not being part of his academic duties. He has always conducted independent research throughout his career and enjoys having papers in the pipeline; a true academic at heart (not by necessity). His focus in the past decade has primarily been on creative entrepreneurship. Topics he has covered include successful entrepreneurship, creativity, innovation, collaboration, crowds, communities, networks & new business models. He not only worked at the university as a professor, but was also a consultant at he same time as publishing papers and wrote a book about what he describes as the “fuzzy firm”: a new future proof network organisation methodology. His diverse career has made him well equipped to partake in the vision for JADS.
The next chapter describes the transition from social science to data science and how both fields can reinforce each other. The second chapter covers insights in the ecosystem of entrepreneurship and data science as a result of his work experience and research. Finally, the third and forth chapter go more in depth about these insights from the field and corresponding applications to JADS.
1. Social Science to Data Science
Using Big Data in Social Studies
Contrary to engineering disciplines there are often many more theories in social sciences according to Arjan: “In physics there is typically one rationale for a phenomena, for example Pascal’s Law for fluid mechanics or the Theory of Relativity for gravitation. On the other hand, in the field of social studies there are often ten unique theories which may also conflict with each other due to the dynamic and complex nature of the issues addressed. Then I sometimes wonder: which theory is the best? This is where I believe big data comes in. One of major problems in social sciences is the replication problem. For example, the results of a conducted study may only hold in 70% of all cases. However, if we combine research and industry data this would become less of a problem because you could then repeatedly test, iterate and improve a theory. This doesn’t necessarily mean we can go back from 10 theories to 1, but I hope we can identify 3 general theories that work really well.”
Subsequently, he elaborates on his source of inspiration by drawing a comparison with primatology: “A great source of inspiration for me is biologist Frans de Waal. Human psychologists gather their information by talking to people. Since we obviously cannot communicate with apes, primatologists are much more dependent on quantitative methods. For example, they measure stress levels, how often apes laugh, in which group they are active, etc. This is very labour intensive work, but you now see machine learning techniques that can automatically analyze video recordings of monkeys. Even though there is great potential, many social studies still use predominantly traditional methods such as surveys and personal assessments. The problem with these methods is that what people say they do is not always the same as what they actually do.”
Have a look at the funny TED-talk from Frans de Waal — born in Den Bosch — below to get an idea of the impact deep learning can have on this type of research.
2. Data Science to Entrepreneurship
Successful Entrepreneurial Teams
In 2012, when Arjan was a Professor of Creative Entrepreneurship at Tilburg University, he contributed to a text book on strategic management: “Gamechangers — How Dutch Companies change the Game”. During our interview he discussed one of the book’s main arguments. As a solo data entrepreneur it is very difficult to succeed: therefore you should always aim for 2 to 3 co-founders. Arjan is not the only one who believes that, for example if you look into YCombinator’s stats you see that 82.1% of all accepted companies in the 2016 batch had two or three founders. They emphasize the need for a co-founder because startups are hard and having a second person to share the load definitely helps moving things forward.
As a solo data entrepreneur it is very difficult to succeed: therefore you should always aim for 2 to 3 co-founders.
The ideal composition of the team, where you’d be optimizing for probability of success, is still somewhat unclear. A venture capitalist would say that you need a team consisting of a designer, business lead and IT specialist. Then the question becomes: how do you make these three parties a cohesive fit? To grow a startup you need different kinds of skills: you need to be administratively strong, have great ideas and be creative. Arjan emphasizes that these competences are in fact opposite things. To be able to manage this paradox is an unique competence. That is why “game-changing” companies are often known for their ability to overcome this issue by bringing together the right mix of people. “What makes the ideal entrepreneurial team?” is an important research question because the answer is crucial to the success of a data-driven ventures.
Another trend Arjan notices is the innovation power of not only universities but also start-ups: “In the field of sports, for example, you now see many data science ventures emerging. At JADS we want to gather such companies around us, be part of it and inspire each other. Moreover, we have a lot more R&D power together.” The Mariënburg campus is the ideal location to create this ecosystem where data-minded organizations can come together in one place.
3. Data Entrepreneurship to JADS
Business Models for Educational Institutions
Although you might not immediately draw the comparison between an university and a company, it turns out there are many similarities. In Arjan’s words: “In essence, JADS is just another start-up. It means we need to raise external capital. After all, for the typical start-up holds: the only path to growth is to raise money and fix problems. Another complexity in our case is the intermingling of numerous business- and organisational models that become dependent on each other. The risk might be high but on the other hand, if everything goes according to plan it’s a unique and hard-to-replicate proposition.”
For the typical start-up holds: the only path to growth is to raise money and fix problems.
You can twist it any way you want but also for a start-up holds that you need to make a profit at the end of the day. Hence, I asked Arjan about JADS’ revenue streams: “In the first place students pay us tuition which is a recurring as well as fairly predictable source of revenue. On top of that research grants is a major source of revenue; you typically talk about 3 to 4 year trajectories. A third channel is professional education, so executive programs for corporations. Lastly there is another business model related to incubation, acceleration and intellectual property. Even though the revenue thereof is limited on the short term, it is crucial to take it seriously from the very beginning because we already put in a significant amount of resources. What I find interesting is that innovation does not only occur at the university anymore, it occurs in the intersection of academia and industry, and there is where we want to be the nucleus of.”
4. JADS to Students
We close every interview with rapid-fire questions for the interviewee. Below you find the answers to three that neatly round up the conversation with excitement and a motto to follow .
What are you most enthusiastic about in Data Science?
“The hard stuff is the easiest, the easiest stuff is the hardest. You can attract the best IT specialist and mathematician but go broke on the simple stuff. Companies come to us with the question: “What will be my business model in five years time from now?”. What intrigues me is that the soft part, which should be the easiest part, is often the hardest part that people get stuck with. All the software packages and mathematical models can be bought, the data integration might be hard but you can do it. However, the simple stuff — how to run a team and make a profitable business — is where people struggle. As a data science company you have to ask yourself what your unique data selling point is. Is it your algorithm that is so specific that it is valuable? Is it your data? Or is it another element? How do you extract that component and sell it? These type of questions are at least as important.”
Data Science: The hard stuff is the easiest, the easiest stuff is the hardest.
What does JADS have in store for the students?
“At the moment we only have 1 program but we hope to start offering 5 to 6 tracks. For example an agro-food track in cooperation with HAS Hogeschool. In this manner students would be able to gain the necessary domain expertise. We are also working on a smart industry track. Additionally, I hope to make JADS a playground where students can integrate their studies with industry partners while running experiments in the wild or on campus. At JADS we hope to offer a variety of domain opportunities from health, crime, smart buildings to the energy market.”
Thinking and doing go hand in hand. If you only think, you never learn.
What are some final words you would like to share?
“Thinking and doing go hand in hand. If you only think you never learn. When you do, you are challenged to think differently. You find yourself in other worlds, other contexts. I would say that 80% of your time you should be doing and 20% of the time you should be thinking and making time to reflect. What I think is very important and what I also try to stimulate is creativity and the ability to ask the right question. In data science, after all, it does not apply that you have data and that you can solve problems with it; there is a level above that which you must ask: What is really the problem?”
Students to JADS
Now it’s is your time to shine 🌟. Please share with us in the comment section what you think of how JADS has progressed, what makes you excited about data science and if you have any words of advice to JADS moving forward?
Thank you Roy for transcribing the interview and co-writing the article.