Matthew Grogan

Teaching R with a focus on tidy, dplyr, and ggplot2

In order to build more familiarity and comfort with R across GIA, I started teaching a short (3 hour) Introduction to R course. I’ve found that the huge steps forward by Hadley and others render a lot of the older guides obsolete.

In the spirit of open-source, I cobbled together my own syllabus. I tried to cover many of the common use cases that data scientists encounter in their daily work.

All the instruction is over at https://github.com/mattgrogan/instruction

Hour 1: Introduction to R

The first hour is an introduction to base R, mostly. It covers some data types, positional and logical indexing, data frames, and tables. The lesson ends with base graphics and a reminder (err, plea!) to follow common coding guidelines.

Hour 1 on GitHub

Hour 2: Data Wrangling

This section is one of my favorites because data scientists often spend a majority of their time acquiring, cleaning, and transforming data. I cannot express how useful it is to have these tools at your disposal.

We used a real-life dataset of grading results on diamonds sent to three different gemological laboratories. We used the tidy package to reshape the data in various ways. Afterwards we jumped into dplyr’s five verbs and then added group_by().

Hour 2 on GitHub

Hour 3: Plotting with ggplot2

Again, as this course is focused on exploratory data analysis, we use ggplot2 to visualize the data in ways that will help guide data scientists as they dive into the data.

I designed the exercises to build up from a simple plot to something relatively complex. The trick here is to be clear about the definitions (geoms, aesthetics, facets, etc.)

Hour 3 on GitHub

Results

Overall we had a great class and I received positive feedback. It is often said that the best way to learn something is to teach it, and I definitely appreciated the effort it took to gather this information into a compelling format.

Back


comments powered by Disqus