Jindřich Čadík: "How to start a successful business and make money"
Hello, I’d like to welcome our guest today, Jaroslav Sixta. Jaroslav is a statistician and educator. He lectures at Unicorn University and the University of Economics in Prague and is also the Vice-Chairman of the Czech Statistical Office. Jaroslav specializes in macroeconomic statistics, national accounts, and advanced statistical methods. In addition to teaching, he participates in numerous research projects, is a co-author of statistical textbooks, and has published over 100 academic articles and studies. Jaroslav, welcome to our podcast.
Good day and hello. 😊
Could you briefly introduce your professional career?
Well, I’m a statistician by profession, and it’s a field I really enjoy. I started focusing on statistics more seriously toward the end of my studies at the University of Economics, when I realized that this discipline needs to be taken seriously and has a lot to offer. Even before completing my master’s degree, I began working at the statistical office as a specialist, focusing on national accounts.
After a short break, I quickly moved on and pursued a doctoral degree. I wrote a dissertation, followed by a habilitation thesis.
Statistics is a field I genuinely enjoy. As you mentioned, I supervise doctoral candidates and diploma theses, and even with my busy schedule, I try to write an article now and then to stay a bit “fresh” in the academic sphere.
And why statistics and data analysis?
That’s a good question. I originally didn’t want to study statistics at all. I chose it at the University of Economics as my second major, with my primary focus being accounting and finance. But back when I was applying, I didn’t get into that. Then I thought, “Wow, this is actually much better.” And that’s precisely what I enjoy. So, it was by coincidence that I found myself doing what I do today.
What brought you to academia and teaching at Unicorn University? What courses do you teach there?
I was inspired by professors at university to pursue an academic career. I consulted and sought advice from them. Basically, I felt it would be a shame to just graduate and not continue contributing somehow. I fell in love with the academic environment, the conferences, and especially the discussions on professional issues. That was exactly what fulfilled me.
Then, of course, students came along, doctoral candidates, diploma students, and that’s how I got into it. I got to Unicorn University through my friend Honza Čadil, who approached me with an offer I couldn’t refuse. It was such a professionally interesting offer that I thought, “This is exactly what I’d like to try.” I hadn’t previously worked with advanced statistical methods or introduced machine learning, and I think I made the right choice because it’s an area that truly fulfills me, broadens my horizons, and allows me to apply some of these fields in practice.
Are those the courses you currently teach?
Exactly. Of course, sometimes you have to teach things you enjoy a bit less. For example, I find basic statistics a bit less exciting. But, as my professor from the University of Economics used to say, teaching the basics is essential to ensure you don’t discourage or demotivate students to the point where they quit not only the school but also the field entirely. On the contrary, it’s rewarding to see people who come straight from high school and you’re introducing them to concepts like mean, variance, or median. Explaining this to them is quite fulfilling in its own way.
I understand. And how do you think teaching statistics and data analysis is evolving with technological advancements?
It’s changing significantly. It’s completely incomparable. When I was studying, we primarily relied on advanced academic works by Professor Hebák, whom I greatly admired. The internet wasn’t as accessible then, and later came its development… Today, the possibilities are just incomparable, whether it’s with the use of AI to help program scripts or find solutions; all you need to do is ask the right question, and you get a script. Then you just tweak it.
But it’s not just about helping with scripts; the field itself is advancing. In the statistical community, things are not entirely free, but they’re at least somewhat accessible. In the past, things took not just a long time but were almost unthinkable to do on a computer. Today, with a better laptop or PC, you can handle and process so much more—it’s really incomparable.
What used to take us half a semester to cover, we now test students on after just two hours. To complete the second part of your question, the distinction between IT, data analysis, and statistics is increasingly blurred. The theoretical part of statistics, of course, will remain and will be the domain of a select few experts. The rest, the mainstream, will handle a lot themselves with the help of AI or someone else. But the differences between programming and statistics in terms of data analysis have blurred significantly.
You’ve already mentioned artificial intelligence and machine learning. How do you see their impact on the future of statistics?
It’s massive! Machine learning, in particular, isn’t entirely new, unlike AI, which is a recent phenomenon of the past few years. Machine learning has been very successful even in statistical practice. For example, at the statistical office, we use certain machine learning methods with great success for classifying unknown objects into specific groups. The machine does it once, learns, and then builds on that.
I wouldn’t say it’s entirely common, but it’s not groundbreaking either. AI, on the other hand, brings something more revolutionary. For instance, you can send a dataset to AI and ask it to: Find all the errors in it, identify suspicious observations, derive some dependency—maybe even a more complex one—to determine the key driver in the dataset.
And what do you think are the biggest trends in statistics and data analysis?
These could be divided into at least two groups. The first trend is in applied statistics in business. Using Python or R, people are solving complex problems that were unimaginable before. This also impacts official statistics to some extent. We’re learning from the best companies in the field and applying their practices in statistics worldwide.
The second area is international progress. In official statistics, there’s a move toward strengthening certain standards, increasing complexity, and placing more emphasis on metadata. Some related fields, like ethics, are also heavily discussed. While I’m not a big fan of that because it detracts from the core, it is somewhat important—for example, interpreting data ethically.
Additionally, with today’s tools, you can craft statistics however you like. You could request any webpage to generate data for you. The question then becomes whether the result is accurate or skewed in any way. How do you maintain trust?
In official statistics, trends are tied to standards, ethics, and legislative coverage. In practical statistics, in business, the focus is on data analysis—making it as precise, fast, and automated as possible. Examples include facial recognition or license plate recognition. Behind these are often simple statistical methods, which, with robust enough data processing, lead to the desired outcome. For instance, speeding tickets are automatically generated when you exceed the speed limit. And, of course, there are more optimistic applications. 😊
What do you think are the biggest challenges students face when studying statistics? And how do you help them overcome these challenges?
So, I’d say the biggest challenge is probably the attitude of students. Over the past five years or so, they’ve somehow gotten used to the idea that they don’t need to study much, that it’ll all just somehow happen. But it doesn’t just happen.
You have unlimited resources available thanks to the internet. And when we teach advanced methods to students, we leave the door open for them to look up things online, to Google… We haven’t allowed AI yet—that’s still a bit much—but maybe we’ll get there someday. However, to be able to use and understand it, you need knowledge. It’s not about memorizing a formula or a definition. It’s about learning how to think, understanding where a potential problem might lie, and figuring out how to approach it. And that’s the hardest part.
So, I’d say it’s partly laziness, lack of preparation, and sometimes even unwillingness. But if you can overcome the laziness, the lack of preparation, and the disappointment of doing poorly on your first test, then statistics as a subject can actually become enjoyable. We have fairly strict tests. On the other hand, we don’t get terrible reviews from students. Most of them eventually say, “Hey, I actually learned something new. That’s nice.” And that’s rewarding.
Let me give an example. We have some mixed groups, including international students. Some of these students are excellent, while others have never even seen Excel before and don’t know how to use it. We try to level the playing field, and the less-prepared students get a bit more work. We strive to ensure fairness for everyone. It’s not easy, but we do our best.
Let’s move on to another topic. Can you tell us about your main responsibilities as the Vice Chairman of the Czech Statistical Office?
My main responsibility is overseeing statistics—that is, the statistics that are published daily. For example, today we released data on inflation, so it’s my responsibility to coordinate activities, help people develop, and ensure adherence to the code of practice and statistical quality. Those are the most important tasks.
I also have top-level management duties, such as strategic planning. For instance, when it comes to deciding whether to implement flash estimates, I prepare everything for the office’s management so they can make an informed decision. So, my role is more about oversight rather than crunching numbers myself. I can afford to delegate that. And that’s what I enjoy about it—sitting over quality reports and emails as a manager, addressing what isn’t working and focusing on more strategic issues.
How do you incorporate your practical experience into teaching at Unicorn University?
I use a lot of examples from areas like price statistics and macroeconomic statistics. For instance, seasonal adjustment is a critical topic. When publishing official statistical data, whether from the Czech Statistical Office or other international institutions, it’s essential to explain to students what seasonal adjustment is and why it’s done. People don’t know much about it, but it’s crucial because you can’t compare unadjusted seasonal data.
We work on this with students—seasonally adjusting GDP using the ARIMA model or experimenting with logistic regression and machine learning for classification tasks. So, it’s very hands-on. We also work with indexes—figuring out what it means if something grows by 3%, then drops by 2%, and calculating the average. We use this a lot.
What are the biggest challenges in the field of statistics in general?
In my opinion, the biggest challenge is maintaining credibility. It’s about ensuring trust in statistics. People shouldn’t just trust the statistics they create themselves—if I may borrow from a classic saying.
We also work on smaller-scale projects, such as refining methods or addressing seasonal adjustments. But the GDP reconstruction project from 1970 was the biggest.
For fiscal policy, it’s just as vital. In this information age, where everyone expects immediate results, markets can react instantly to statistical releases. It’s extremely important.
In business, the approach is a bit different and more stable. If you run a company, for example, and need to determine the percentage of fraudulent loan or insurance claims, the approach is less media-driven. In public administration, everything is under greater media scrutiny. For businesses, however, it’s more about deciding on an investment—whether it’s profitable or not. Both contexts require reliable data, but they approach it differently.
On the supply side, however, I’m less optimistic. There’s a significant challenge in finding people willing to work in this field. It’s not that they’re uninterested; rather, the financial compensation isn’t always adequate. This makes it hard to fill positions, not just at the statistical office but in many state institutions.
