Q&A with Dr. Jakub Stoklosa
Dr. Jakub Stoklosa is a Senior Lecturer at the UNSW School of Mathematics and Statistics and will be familiar to those who have been in his MATH1041 and MATH2801 courses. We recently conducted a written Q&A with him to learn more about his life, career, and research interests as a statistician. We even snuck in a few fun questions at the end (you’ll be glad we did). Read on for the full Q&A.
Organised by David Bee Olmedo (with additional questions provided by Jack Stephens)
Background Questions
1. How many years have you been at UNSW? Of those, how many have you spent lecturing? Where were you before UNSW?
I have been at UNSW for 12 years now. Officially, I have been lecturing since the start of 2017. Prior to 2012, I did my undergraduate degree and obtained my PhD from the University of Melbourne.
2. What courses do you currently teach?
I teach MATH1041: Statistics for Life and Social Sciences, and MATH2801: Theory of Statistics.
3. What did you do your PhD on? Any surprising findings or experiences which you would like to share?
My PhD thesis developed a range of modern statistical methods for analyzing capture-recapture data. Capture-recapture is a sampling technique, often used in ecology to help estimate population demographics, like the unknown population size of animals or their survival rates.
One of my PhD projects analysed data on Little Penguins. There is a massive colony of them on Phillip Island (south of Victoria). Little Penguins will either nest in natural or artificial (man made) burrows when they are not at sea. My co-supervisor at the time was interested in seeing if Little Penguins had a preference for switching to different nesting types over time. Using some statistical modelling we found that Little Penguins generally preferred natural burrows over artificial burrows, and would more likely switch to natural burrows.
4. What are your current research interests?
I still do a lot of work on capture-recapture methodology and abundance estimation. I also work on species distribution modelling, model selection, and measurement error modelling. The latter is a statistical technique that is used for correcting imprecise measurements of observed variables, such as, temperature recordings.
Main Questions
1. What inspired you to pursue a career in statistics?
I have always enjoyed all forms of mathematics (pure, applied and stats) throughout my undergraduate, but I felt that statistics was more “hands on”. Also, my major was in financial mathematics, so I needed to do lots of probability and statistics subjects, which I really enjoyed.
1.1 Follow-up: Did you ever see yourself lecturing in statistics?
Nope. My plan was always to head into the industry (private sector), but after my Honours year, I decided to do further research. Several years later, here we are…
1.2 Follow-up: What would you say is the most satisfying part of your job?
In terms of teaching, I would say the student engagement is the best part. In terms of research, collaborating with other researchers from various disciplines
2. What word of advice would you give any students who are currently majoring in statistics?
Always brush up and revisit your linear algebra and calculus. These topics will be helpful for further studies in stats.
2.1 Follow-up: To a person with no statistics background, what would you say are the top two reasons statistics is essential in today’s world?
I guess uncertainty is all around us and it can play a big role in our everyday lives. Statistics/probability is one way of understanding uncertainty through data.
3. Which research project has given you the most fulfilment to do? Any findings you would like more people to know about?
I really enjoyed my first Post doc project with Prof David Warton. We developed a new model selection criterion for multivariate data with many predictor variables. The advantage of the proposed method was the computational cost which greatly improved the model fitting speed. In summary, we managed to cut down our data analysis time from ~2 hours to ~22 seconds!
4. What would you say is the most interesting thing you’ve worked on in your career so far?
In terms of statistics and methodology, I would probably say the paper above, that was good fun. In terms of application, I was involved in a long-term project which focused on the genetic rescue and translocation of Mountain Pygmy Possums in Mount Buller, Victoria.
5. Have you done any group-based research? Do you have any tips for students struggling with group assignments?
Yes, for sure. In fact, all my research has had at least one collaborator. In terms of advice, I would say start as early as you can. It gives you more time to find your strengths and how you can contribute.
6. How crucial would you say R (and equally RStudio) is to your day-to-day work?
I would say it is very essential. I would also recommend learning RMarkdown.
6.1 Follow-up: Do you recommend all statistics majors master it as early as they can?
Yes, for sure. Get as comfortable as you can with R/RStudio.
6.2 Related: Are there any other software packages that are perhaps not as well known yet very useful in your work?
I have heard good things about Python. TensorFlow might also be a good one to learn about.
7. In your opinion, is there a significant distinction between statistics and mathematics? If so, what would you say distinguishes statistics as a separate stream of study?
This is an age-old question that I sometimes think about. There certainly are some differences but I will always treat statistics as a branch from the mathematics tree.
Fun Questions
1. Online teaching. Love it or hate it?
Love it. But it does take some getting used too.
2. Do you have any hobbies you don’t mind sharing?
I am a big NBA basketball fan. I played quite a bit when I was younger. I also did a lot of rock-climbing indoor and outdoor) during my teens. I occasionally do a bit of bouldering when I can.
3. Is there any particular result/theorem/etc. which you have utilised hundreds of times yet still need to look up every now and then?
Yes, the AIC! Akaia Information Criteria (or AIC) is a well-known and commonly used model selection criterion – it is a tool for selecting the “best” model from a set of candidate models. The formula is quite simple, it goes AIC = - 2log-likelihood value + 2(no. of parameters in your model), David, you will recall what the likelihood value is from MATH2801. For some reason though, I always forgot whether we add or subtract the 2 in the AIC formula…
4. Do you ever find yourself calculating statistics of everyday things/incidences?
Yes, all the time. As a statistician/mathematician you tend to think in terms of probability and risk. Sometimes this works in your favour and other times not so much.
5. Do you know any good statistics jokes?
I have a joke about P-values but it isn’t significant.
6. Blackboard or whiteboard?
Whiteboard for sure.
7. What’s something people might not know about you?
I went to the same high school as Chris Hemsworth (aka Thor). In fact, my older brother was good mates with him throughout high school.
8. What’s your favourite music to listen to?
I guess my taste for music has changed over the years, but I’ve always stayed true to trance.
9. Favourite movie?
A South Korean movie called “Oldboy”.