tjmahr / MadR_Pipelines

Slides and materials for my talk to the Madison R Users Group

Let the Data Flow: Pipelines in R with dplyr and magrittr

Abstract

Pipelines were the best thing to happen in R in 2014. They let us transform messy, inside-out code like sort(unique(round(xs, 2))) into a clear chain of transformations like xs %>% round(2) %>% unique %>% sort. In this talk I lead a tutorial on how to use pipelines for data-cleaning, transformation and presentation with the packages magrittr and dplyr. For beginners, I also review some of the essential R functions to make the most of pipelines.

Tristan is a PhD student in Communication Sciences and Disorders. He uses in R in the Learning To Talk lab to model eye-tracking and speech perception data. @tjmahr, github.com/tjmahr.

Slides

I prepared three sets of slides:

• Pipelines in R
• dplyr: verbs for manipulating data-frames
• Making pretty regression tables with pipes

Resources

• magrittr vignette
• RStudio Data-Wrangling Cheatsheet
• Core R vocabulary
• Awesome R
• Pipelines for Data Analysis (dpylr/magrittr talk by Hadley Wickham)
• Best Practices for Scientific Computing
• Data Science on the Command Line
• Unix Commands for Data Science

Packages

• magrittr for pipelines
• dplyr for data-frame functions
• broom
• stringr for string manipulation functions
• pipeR an alternative pipeline package (that I haven't tried yet).

License

Obviously, the GPL-2 license applies only to the code and words I wrote, which are in the .Rpres and .md files and are reproduced with markup in the .html files.