IDST 290: Data Science for COVID-19, Fall Semester 2020


Continuing to traverse the new, uncharted territory of a mid-pandemic world, the University of NC at Chapel Hill requested course proposals on the SARS-CoV-2 epidemic from the perspectives of different academic fields. It is with this direction that myself and three of my colleagues at UNC -- Jan Hannig, Serhan Ziya, and Richard Smith -- created a course covering how a data scientist might approach the problems that arise in a global catastrophe like the COVID-19 pandemic. In addition to exposing our students to an impressive roster of speakers, this course aimed to motivate first-year students to explore their interest in data and learn about ways that they can tailor their Carolina experience to position them well for careers that capitalize on these interests. Notable guest speakers for this class were Sir Dr. David Spiegelhalter, Dr. Vukosi Marivate, and Dr. Christl Donnelly. For the full roster of speakers, and a break-down of the course in general, take a look at our syllabus.

One of the most exciting and thought-engaging elements of this course was the podcast series hosted by the COVID Investigations program. These podcasts recorded in-depth interviews with scientists who are leading the global effort to fight SARS-CoV-2. I would highly encourage anyone interested in how the world responds to a pandemic to take a listen! My personal favorite is the one with Dr. Myron Cohen.
  • Podcasts
  • Videos of lectures, except for the one hosted by Dr. Donnelly (who wished to limit her public exposure), are posted here. My favorites were the talk by Peter Frazier on Cornell's (successful) COVID response and the talk by David Spiegelhalter.
  • Lecture Recordings
  • The course evaluations for my class are available here.

    Statement about Diversity in the Statistics Classroom: We celebrated diversity in our virtual classroom in two ways. First, our guest speakers were intentionally gathered from all over the world; we heard talks from South Korea, England, and South Africa -- as well as local scholars from the US. This international roster of speakers mirrored what was a fully international classroom; we had over 100 students hailing from 12 countries taking the class in 12 different timezones. Second, many of our podcasts and subsequent podcast talks tackled issues of diversity and identity. We used these "breaks from data science" as a way to get students to engage with these topics while also learning more about each other.

    STOR 155: Introduction to Data Analysis, Summer Session I 2020


    In the wake of the COVID-19 pandemic, I lost my internship at the Mayo Clinic and in turn signed onto a teaching position for UNC's Summer Session I. Due to health concerns related to the pandemic, all teaching was remote. I designed my course in a way that I believed would make the difficult and sudden transition to remote learning as painless as possible for my students, while still demanding diligence and genuine mastery of the material.

    The foundation of my course was a daily routine of a 1 hour live lecture, a shorter video supplementary lecture, and an online homework assignment using the online portal WebAssign. Myself, and my team of four instructor's assistants, spread out over 38 office hours throughout the week every week. Since I wanted to maximize the amount of practice available to my students, much of my work this summer involved developing WebAssign questions in Perl. For those interested in learning Perl code for WebAssign development, I have attached three useful examples for developing your own questions. I am quite a fan of questions where you simply give a student a set of data and have them perform analysis and inference. The first of these examples shows a way in which one can do this, and how to use continuous distributions in Perl. The second outlines using discrete probability distributions in Perl, and includes my own implementation of the Negative Binomial CDF. The third is a diversity question (see statement below). The fourth gives 5 different sets of data and asks students which pairs satisfy the conditions for least squares regression.

    The course evaluations for my class are available here and here.

    For the purposes of your own teaching, I encourage to use any of the following materials that I have developed:

    STOR 155 Materials: Statement about Diversity in the Statistics Classroom: While leveraging the diverse perspectives of students in the classroom is one of UNC's pillars of teaching, this can be a tricky thing to do in a quantitative class. In an attempt to integrate the many exciting perspectives brought by my students into our learning, I incorporated each of my students' worldviews into our homework/exam questions and our in-class examples. See Webassign Perl Code Example 3 for a question I used on our first homework assignment to garner this information from my students. This semester was certainly a trial run of this idea; I can say now that I am very pleased with the results! Besides making my students feel more included, it was quite enjoyable to learn more about what is important to each of them.

    STOR 565: Machine Learning, Spring 2020


    The following are the computing assignments (CAs) used for the course I am taught with Dr. Andrew Nobel during Spring 2020. These assignments are written by myself, with the occasional (shameless) stealing of exercises from my collegues Kelly Bodwin, Iain Carmichael, and James Wilson. The course evaluations of our class are available here.

    Computer Assignments: Update as of March 18th, 2020: Due to the Coronavirus pandemic, submission of all Computer Assignments from CA08 and onward were submitted using the online portal gradescope. Furthermore, all instruction was quickly adapted to be remote.

    Final Project: Teaching while dealing with a global pandemic was a tremulous affair. Dr. Andrew Nobel and I were very careful to balance the work load required to really learn the material with what we believed to be a reasonable work load to give to students in distress. In the end, I wrote two projects: the full draft, which would have been given under regular circumstances, and the second draft, which we believed was more manageable for our students.