Community Spotlight #1: Keith Lohse
November 10, 2022
Keith Lohse is a Clinical Associate Professor of Physical Therapy and Neurology at the Washington University School of Medicine in Saint Louis. He is accredited as a professional statistician (PStat) by the American Statistical Association. He has expertise in multivariate statistics, signal processing, and longitudinal/time-series data. He has recently been learning about machine learning, especially unsupervised and reinforcement learning algorithms. He enjoys problem solving, developing new data visualization skills, and the challenge of translating complex data to a diverse audience.
Tell us about your background.
I started college at Idaho State University very unsure about what I wanted to do with my career. I really liked cartoons and graphic design, but I thought it would be more pragmatic to study science. I looked at the undergraduate course catalogue and saw that psychology was the science with the fewest required classes, so I signed up for psychology in the hope of exploring the most electives. Although I did explore a lot of fun electives in linguistics and philosophy, I really got hooked by taking a psychology-specific statistics course in my first semester. From there I went on to take more advanced statistics in the math department and that set me on a path of wanting to learn as much as I could about measurement, design, and analysis within the world of rehabilitation.
What sparked your interest in learning data science?
From my first class, I was really fascinated by statistics and how people used inferential statistics to test hypotheses in science. As I learned more about just how hard that is to do, I also learned about how trials can give discrepant results, how many people you need to gather data on, and the difficulties of integrating data from different sources. I still know vastly more about statistics than I do about data science proper, but more and more of my work is focused on making rehabilitation data Findable, Accessible, Interoperable, and Reusable (FAIR).
What are you most excited about ReproRehab?
I am excited to help build this network of researchers and cultivate data skills in rehabilitation science. I am grateful for this opportunity to give back to my community and create educational materials that are openly available to anyone who is interested all over the world.
What advice would you give to those who wish to follow a similar path?
Students often tell me that they don’t have a “math brain” or just “don’t get programming”. But this seems like the only domain in which students feel this is a reasonable excuse; students never lament their lack of “biology brain” or how they just “don’t get words”. In my experience, students think that I am good at math or programming because of some natural aptitude, when the reality is that I have these skills because I SURVIVED math and programming.
I think it is important to remember that learning is a two-way street. As a learner you need to bring motivation and diligence to a problem, but if you still don’t “get it”, maybe the material is not being presented in the best way for you. Math never really clicks for me until I find the need to apply it to a specific problem and can use it in a context I understand. If you feel like you can’t “get it” despite trying your best, seek out and explore new environments for learning. Try to find a problem that is important to you and commit yourself to mastering the tools you need to solve it.
Anything you are passionate about outside of ReproRehab?
I love spending time with my wife, Emma, and our two dogs, Olive and Moose!
I also enjoy hiking, going on runs around Saint Louis, and exploring new and interesting music. There are a lot of great genres of music out there, but I love rock operas - Blind Guardian’s “Nightfall on Middle Earth” being my favorite, narrowly beating Rush’s “2112”!
Anything else you would to share? (website/github/podcast)
People can always checkout my personal website, which includes links to my GitHub repositories and YouTube channel: https://sites.google.com/site/lohsekr/
