R Workshop Notes

• Software and Coding Basics
• Data Management
• Loops
• Writing Functions
• Statistics
  • Some Statistics: Descriptives and Hypotheses
    
  • Linear Models
    
• Graphics
  • Graphics
    

The User Interface

May I see the menu, please?

Basics:

• Working in RStudio (recommended)
• Working with R Console

Additional topics:

• Using R on Linstat: linux command line usage
• Using RStudio with Linstat: terminal usage

Data: Objects in the Environment

Everything is data

• Data Wrangling with R
• Some Data Wrangling References

Reading Data into R

Data has to come from somewhere

• Text data (see Data Wrangling)
• Foreign formats (SAS, Stata, Excel, etc.): binary.r

Conditional and Repeated Tasks

• Loops

Writing functions

• Writing Functions

  • plotmethods.r: methods for logical and character vectors

  • scale.data.frame.r: a scale() method for data frames

  • mean.data.frame.r: a mean() method for data frames

  • Dot-and-whisker plots: means and standard deviation, using substitute() dotandwhisker.r

  • Replace a value: Anonymous functions, and using lapply() writing a modified function: replacevalues.r

  • Using tapply(), extracting from lists "[[ ]]": Histograms by group with graphics

  • A graphics function to add a normal curve to a histogram: histnorm.r

Some Statistics: Descriptives and Hypotheses

Evidence and probability

• Frequencies and crosstabs: simplestats.r
• Means, t-tests, One-way, correlation: morestats.r

More Statistics: Linear Models

• Regression Formulae and Model Building
• Simple models: simplemodels.r
• The language of models: linearmodels.r
• Parameter Translation

Graphics

Seeing is believing

• Three graphics packages: graphicssystems.r

  There are three commonly used graphics packages in R: graphics, lattice, and ggplot2.

  • graphics is the oldest, and is often called 'default graphics' or 'base graphics.' This collection of functions allows you to build images interactively, layer by layer.
  • lattice was designed to emulate S's trellis graphics, and makes it especially easy to render comparable graphs of subsets in a data set. It returns objects in your workspace that can be further queried or modified. However, layering graph elements is a bit awkward in lattice.
  • ggplot2 was designed as an implementation of Wilkinson's Grammer of Graphics and has some of the good qualities of both lattice and graphics.
  • Many other packages also include graphing functions.

graphics, the default graphics package: graphics.r

• The Plot Gallery
• Useful Graphics options
• Side by side histograms (and a comparison of graphics versus lattice)

Lattice, or stats revisited: lattice.r

ggplot2 examples: Additive and Interaction Models, layered in facets

Exercises 6 (lattice)

Answers 6

Other Topics

Really? You want more?

• Building packages

  I’d start with Karl Broman’s Package Primer