Albert Y. Kim, Jordan Moody, Ziwei "Crystal" Zang, & Starry Zhou (Smith College)
Wednesday May 15, 1:00 pm – 4:30 pm
Thursday May 16, 8:30 am – 12:00 pm
We propose an approach to learning, teaching, and performing statistical inference using tidyverse data science tools. We argue that this tidyverse-centric approach is both 1) feasible since the tidyverse is more intuitive for new R coders to learn than base R and 2) valuable since on our proposed path to statistical inference students will also learn to use data science tools applicable beyond the classroom.
We’ll start with “just enough” data visualization & wrangling with the ggplot2 & dplyr packages to equip students with the necessary computational tools for the journey. Using these tools, we’ll cover both explanatory & predictive regression modeling and the infer package: a new package that makes statistical inference “tidy” and transparent. A combination of pen/paper exercises, “tactile” sampling/resampling simulations, and a live experiment whose data will be analyzed on-the-fly will keep participants engaged.