鈥淎s an experienced advisor, I wanted a hands-on program to learn modern AI techniques.鈥�

Philander van Hien

With over 25 years of experience in the data domain, Philander van Hien joined the Mastering Data & AI program at EAISI Academy. Daniel Kapitan spoke with him about what the program brought him. In addition to applying AI in data migrations, they also discussed his passion for flying and how he developed an app to predict low cloud cover along flight routes.

DK: To start off, how did you come across EAISI Academy?

鈥淲hen I started the program, I had more than 25 years of experience guiding major data migration projects in the financial sector. Savings, investments, leasing, mortgages鈥攜ou name it, I鈥檝e been deeply involved in this world for a long time. Naturally, I was seeing more and more applications of machine learning in this area. I was seeking a way to establish a solid foundation, both in theory and in practical application. That鈥檚 how I found the Mastering Data & AI program.鈥�

DK: How did you experience the structure of the program?

鈥淚t aligned perfectly with what I was looking for. During the first module, Introduction, I established a solid foundation and took my first steps into AI. It was important to me that the concepts and principles were explained in a structured way. This enabled me to leverage my extensive experience with modern AI applications.

In the second module, Applications, the importance of structured development using the CRISP-DM method became clear to me. I collaborated with three other participants on a case study about disability insurance, a topic closely related to my field. What I appreciated most was that the program went beyond teaching 鈥榯echnical tricks.鈥� In the business simulation, where we had to solve a specific problem, I experienced what it鈥檚 like to work on a data and AI project as a team. Thanks to that experience, I now have a much better grasp of how to practically structure and manage projects. And it鈥檚 not just about the technical or analytical work鈥攊t鈥檚 also about creating space within your team for more conceptual discussions and gaining a deep understanding of the business problem.鈥�

DK: Looking back on the program, what advice would you give to future participants?

鈥淭he program helped me grow in two roles: as an analytics translator and as a data scientist. I鈥檇 tell new participants that the choice between those two roles isn鈥檛 as black and white as it may seem at first.

Additionally, planning and effective time management are crucial. The lectures, reading materials, and projects took more time than I initially expected. Luckily, I found an effective approach by scheduling fixed study times for each component. The online learning platform supported this with its clear structure. I even utilized generative AI to establish a routine with daily microlearning tasks.

For me, the challenge-based learning approach was especially effective. Rather than passively listening to lectures, I was really challenged to actively apply what I鈥檇 learned to real-world problems. That worked very well for me鈥攁nd likely for other experienced professionals as well. I was able to immediately apply what I learned both in my work and my hobby.鈥�

DK: You had a very unique Mastering project during THE FINAL MODULE of the program that tied into your hobby. Can you tell us more about it?

鈥淵es, with pleasure! As a hobby, I鈥檓 a private pilot and fly according to VFR rules鈥擵isual Flight Rules. This means flying based on visual reference. Two key meteorological factors influencing this are visibility (the distance at which objects can be seen) and cloud base (the height of the clouds). As a VFR pilot, you must always be able to see the ground and maintain a safe distance from the clouds.

The problem I sought to address was the scarcity of detailed weather data for private pilots. Most services provide general forecasts for large regions, but they often overlook the specific details required for individual flight routes. That led to my idea for the Virtual Co-Pilot project: a tool that generates real-time forecasts for visibility and a cloud base tailored to your planned flight path.

For this project, I employed the CRISP-DM methodology, which I learned during the program. It鈥檚 a structured approach to data science and AI projects, working iteratively: start with a basic model, test it, improve it, and repeat. It鈥檚 like following a recipe that you keep refining. And since it鈥檚 a standardized approach, you can apply it to any project, in any industry.

A key aspect of my project was gathering the right data sources. Initially, I worked with a commercial database, but it suddenly moved behind a paywall. Iterating with your model becomes expensive fast if you have to pay for data. I switched to open-source platforms, such as the KNMI (Netherlands), KMI (Belgium), and the UK鈥檚 Met Office. The big advantage of open data is that it鈥檚 accessible via APIs and free to use. This perfectly illustrates the value of open data鈥攊t enables innovation without huge upfront costs.

Of course, I faced some challenges. I鈥檓 not a hardcore IT specialist, so connecting to APIs and processing the data took extra time. Luckily, pilots can call the weather service鈥攖hose experts gave me a helpful nudge in the right direction. And since I had to build this project alongside my regular work, time management was crucial. I often worked into the evening to meet the program鈥檚 deadlines.

The result is a working website: , which is freely accessible to everyone. As a hobby pilot, I now use the tool for every flight preparation. The mobile app is also expected to launch in the app stores soon. One thing I鈥檝e learned is to keep a backlog of ideas and improvements. A project like this is never really 鈥榝inished鈥欌�攖here are always new features to add or tweaks to make.

What I find most valuable about this project is that it effectively demonstrates how data science and AI can be applied to a real-world problem from your own life or work. Because I fly myself, I could clearly define the need and the desired solution. By utilizing open data, I鈥檝e developed a sustainable solution that is independent of commercial providers. That, to me, is the power of open data: it democratizes innovation and makes it possible to build solutions that otherwise might not be feasible.鈥�