RStudio First Impressions: Graphapalooza
April 2018 (perspective of an assistant professor)
Almost five years ago I wrote about my first impressions of using IPython Notebooks (now called Jupyter Notebooks).
In a similar vein, I've been meaning to learn the modern R programming ecosystem and the RStudio IDE for quite a while now. A good forcing function came recently when I had to make a new set of graphs for a research presentation about my Python Tutor project. I decided to try to use R again and capture my first impressions in real time.
I had already painstakingly cleaned my data beforehand so I just needed to make graphs. Thus I embarked on a two-hour graphapalooza and narrated my thought process. The last time I programmed in R was over 10 years ago, and I've never used any of the modern libraries. Thus, these videos serve as an informal N=1 user test of creating graphs using RStudio and parts of the modern R ecosystem.
Check them out! (full YouTube video playlist)
After finishing these two videos, I ended up tweaking my graphs more and learned additional features like being able to execute individual blocks of code directly from the editor using a keyboard shortcut. That prevented me from having to constantly fight with the kinda-janky REPL.
Overall I feel pretty good about using R in the future for creating graphs and doing simple data transformations once my data is cleaned to a tidy format. But I'll still stick with Python for the bulk of my data wrangling and pre-processing needs; I'm much more familiar with that language, and I already have a lot of data processing code written in it.
UPDATE: exactly one month later (May 8, 2018) I returned to RStudio to make more graphs, so I recorded four more videos ...