Visualization & Reporting

Submit Attendance: link

Files for today: functions | base plots | ggplot2 | quarto | solutions

Functions and Control Flow

  • Use the function() function to create your own function
  • Functions have:
    • Formals, which are the list of arguments that control how you call the function
    • A body, which returns the last evaluated expression or whater you put in return()
    • An environment, which determines scoping (ie, how the function finds the values associated with named objects)
  • There are some R-specific things to be aware of
    • Lazy evaluation
    • Things that appear to have the same name, eg fun(x=x)
    • Anonymous “lambda” functions
  • Control Flow
    • if() {...} else {...}, and ifelse()
    • with assignment
    • loops via foreach() and while()

Base Plots

  • R has excellent plotting capabilities, but often requires lots of code to fully customize the plot
  • Base plots are static
    • No resizing or zooming
    • Adding elements to a plot is done by layering on top, not modifying what exists
  • You’ll often want to change:
    • The color col, point character pch, fill color fill, line width lwd, or size via character expansion cex
    • The title main, horizonal axis label xlab, and vertical axis label ylab
    • Margins via par(mar=c(...))
    • Any many other features, several highlighted in these [article1, article2]
  • Add elements to your plot with:
    • lines(), curve(), text(), or points()
    • or customize the axis()
    • and add a legend()
  • To combine multiple plots:
    • Use par(mfrow=c(x,y))
    • Control outer margins with par(oma=c(...))
    • Place a title via title()
    • layout() offer richer arrangements of subplots

ggplot2

  • An introduction:
    • The ggplot2 package is Hadley Wickmham’s implementation of a Grammar of Graphics described in Leland Wilkinson 2005 book
    • It is immensely popular, with re-implementations in Python and Julia
    • It is powerful and expressive method of creating graphics, and well documented with a website, book, and many other resources
  • Every ggplot involves
    • A dataset usually via ggplot(data)
    • Aesthetic mappings from variables in the dataset to plot elements, eg, aes(x=height, y=weight)
    • One or more geometric objects (eg, points in a scatterplot via geom_point())
  • You’ll often want to annotate basic elements with labs() such as x, y, title, and subtitle
  • More advanced aspects of the plotting system involve
    • A statistical transformation of the data
    • A scale to control the mapping from data to aesthetic attributes
    • A coordinate system
    • Facets, for creating small multiples of plots
  • The theme system provides fine control over the non-data elements of the plot

Quarto

  • Quarto is an open-source scientified and technical publishing system created by the team at Posit
  • A Quarto document
    • Can by a Jupyter Notebook or a plain text .qmd file with a YAML header
    • Consisting of Pandoc markdown and code chunks
    • Renders to multiple output formats including PDF, MS Word, HTML
    • Enables reproducible research and literate programming
    • Will likely be the mode via which you compose your homework assignments
  • Pandoc markdown syntax:
    • Is quite simple and the basics can be learned in an hour
    • Includes Latex math equations, citations, cross references, and more
  • Code chunks:
    • Begin and end with three back ticks `
    • Support a variety of languages, including R, Python, Julia, C++, BASH