For the past seven years, I have taught statistics and data science as an instructor of courses, labs, and workshops. Students describe me as enthusiastic and accessible and I most enjoy providing powerful examples to make complex methods more tangible. My goal is to train audiences in becoming critical consumers and producers of numbers, and to motivate anyone to learn basic methods of scientific inquiry. I am most proud of having taught Python workshops that introduce students and faculty to computational methods (see workshop links below.)
For a full teaching portfolio and student evaluations, please send me an email.
Python Workshops:
· Exploratory Data Analytics & Visual Analytics (2024)
· Network Methods for an Interconnected World (2024)
· Advanced Models Using Text (2023)
· Introduction to Network Analysis in Python using networkx (2022)
· Text Mining in Python (2019, 2020)
· Webscraping and APIs Using Python (2019, 2020)
· Introduction to Python (2018, 2018, 2019, 2020)
Workshops were taught as part of the Workshop in Methods Series at the Social Science Research Commons (Indiana University), the Network Science Institute (Indiana University) , and the Massive Data Institute (Georgetown University).
Graduate Courses (Lab Instructor):
· Categorical Data Analysis (2023)
· ICPSR: Data Science and Text Analysis (2020)
· Statistical Techniques in Sociology I (2017, 2018, 2022)
Labs were taught in the Sociology Department at Indiana University and at the University of Michigan as part of the Inter-University Consortium for Political and Social Research (ICPSR) Summer Program.
Undergraduate Courses:
· Understanding Social Data (2018)
Associate Instructor, Sociology Department at Indiana University.
For a demonstration of how I try to make complex methods more tangible, consider this web app, which I coded in R (Shiny package). It allows students to examine the nonlinear dynamics captured by logistic regression models by playing around with the model parameters and observing the effect on the predicted probabilities. The typical logistic regression output of any statistical software is logged-odd ratios, which are difficult to interpret and give a false impression of linearity. To fully examine model results, researchers need to work with predicted probabilities (e.g., Average Marginal Effects). Yet unlike logged-odd ratios, predicted probabilities do not scale linearly, which students can examine through the app.