Link to the slides

http://tinyurl.com/zzlzs2u

Advantages of building an R package

• Use functions and turn hundreds lines of code to a single or few lines of code.

• Orgainze your R script, data, and documentation in a well-defined structure.

• Collaborate and share your methods with others. Editing history traceable via git control.

• Use ‘R CMD check’ to validate your R script.

• Publishable work

R package structure

A minimal package should include:

• R functions (saved in .R files)

• Metadata (DESCRIPTION file)

• Namespace (NAMESPACE file generated via roxygen2; connect your package to others)

• Documentation (generated via roxygen2; help page)

• Vignette (user guide)

Writing R functions can save your time down the road

How to write a function

Your_Function_Name <- function( Your_Variables ){
  #Do something here
}

fecundity <- function(L){
  y <- 0.002*(L^3)
  return(y)
}

fecundity(36)

## [1] 93.312

How to write a function

Fecundity = a x Length ^b

fecundity <- function(L, a, b){
  y <- a*(L^b)
  return(y)
}

fecundity(L=36, a=0.002, b=3)

## [1] 93.312

fecundity(L=36, a=0.2, b=3)

## [1] 9331.2

Make your function useful

Use conditional statement to choose an analytical method

myFunction <- function(method="A"){
  
  if(method=="A"){
    #Analyze data in this way
  
  }else if(method=="B"){
    #Analyze data in that way
  
  }else if(method=="C"){
    #Analzye data in that way
    
  }
}

Start an R package with git version control

Your R package is built in your computer; a copy can be saved in the cloud. They’re syncronized via Git.

Start an R package

• Install R, RStudio, and Git in your computer

• Create a Github account www.github.com

• Create a new repository / Copy the repository’s link

• Create an R project from RStudio
  1. Open a new project
  2. Select version control
  3. Select Git
  4. Paste URL

• Install library devtools and roxygen2

• Set up R configuration

R configuration setup

1. Select Build and then Configure Build Tools on menu bar

2. Select Package in the pop-up window

3. Check and configure Roxygen options

4. Create folders R and man

5. Create a text file named DESCRIPTION

Edit DESCRIPTION file

Include the following information:

Package: MyRpackage
Version: 1.0
Title: This is title of MyRpackage
Author: Alex Chen
Maintainer: Alex Chen <chen.1735@osu.edu>
Description: A brief description about MyRpackage
Depends: R (>= 3.1.0)
License: GPL (>= 2)
Imports: [Put other packages here if needed]

More details about DESCRIPTION

R official site

by Hadley Wickham

Write R function files

Create function files (e.g., fecundity.R)
and save them in the R folder

#' A title of your function
#'
#' A short description about the function.
#' @param L fish length
#' @param a constant a
#' @param b constant b
#' @return This function returns fish fecundity
#' @export
#' @examples
#' fecundity(36, 0.2, 3)

fecundity <- function(L, a, b){
  y <- a*(L^b)
  return(y)
}

We use roxygen2 to generate functions’ documentation.

Build your package in R

• Click Build & Reload under Build menu

• Fix any issue if build not successful

• Make sure to rebuild if any change was made

• Try ?fecundiy to see if help page shows up

Sync with Github repository

• Make sure the Git program was installed.

• Click Commit under Git menu

• Check files and add commit message and then hit commit

• Close commit message window

• Click Push button to upload files to Github repository

• Go to your Github site and see if files are updated

Exercise

• Go ahead and create a second R function

• Rebuild your package locally, and then upload the file to your Github

• Install the package of a person next to you. Use the following code:

library(devtools)

install_github("ACCOUNT/XPACKAGE")

Handle identical function names

• Import the library: library(XPACKAGE)

• Type ?fecundity in R console (There should be two fecundity functions associated with different packages)

• Type ?XPACKAGE::fecundity to see the help page

• Click Index at the bottom to see the second function

• Run the second function to see if it works

Further resources

One-page tutorial (by Hilary Parker)

R package primer (by Karl Broman)

R packages (by Hadley Wickham)

R official document

_~

Happy coding!