Jonas Kersulis, PhD

Power systems, data science, interactive visualization, teaching

I completed a PhD in Electrical Engineering at the University of Michigan in 2020. I am currently assisting Professor Raj Nadakuditi in the development of interactive data science teaching material, while searching for full-time employment.

Power systems research

During my PhD I worked with Dr. Ian Hiskens to study various impacts of wind fluctuations on transmission and subtransmission networks. I also studied the graph structure of transmission networks, electric load profile time series, and tap-changing transformer behavior in the presence of renewable fluctuations. Check out my research page to find out more, or email me if you would like to discuss.

Data science

My current work for Professor Raj Nadakuditi involves editing, testing, and expanding interactive teaching material for a course on computational data science. In its modern form, the course has been taught to hundreds of students at both the University of Michigan and MIT. I have been a member of the Graduate Student Instructor (GSI) team for the course several times. This experience has made me deeply familiar with numerous data science theoretical concepts and algorithm implementations.

Interactive visualization

I have always been passionate about visualizing data. Over the course of my research I encountered a great deal of data, often quite complex. In some cases a static visualization cannot capture what I would like to see; something dynamic is required. It is at these times that I find the greatest opportunities to leverage interactivity and motion to help myself and others explore and experience the data. The following posts introduce a few examples of dynamic or interactive visualizations I’ve crafted:

My passion for rich, interactive visualizations is closely tied to my interest in teaching.

Graduate and undergraduate teaching

I love making sense of complicated ideas and communicating them to others. I give presentations, of course, but my favorite communication tools are visualization and code, as illustrated by my teaching projects. This approach is well-suited to EECS 551/598 (Mathematical Methods for Signal Processing) at the University of Michigan. This course is rapidly developing into a large-scale, multi-university computational data science course.