Introducing R

R is a system for data manipulation, calculation, and graphics. It provides:

Facilities for data handling and storage
A large collection of tools for data analysis
Graphical facilities for data analysis and display
A simple but powerful programming language

R is often described as an environment for working with data. This is in contrast to a package which is a collection of very specific tools. R is not strictly a statistics system but a system that provides many classical and modern statistical procedures as part of a broader data-analysis tool. This is an important difference between R and other statistical systems. In R a statistical analysis is usually performed as a series of steps with intermediate results being stored in objects. Systems such as SPSS and SAS provide copious output from (e.g.) a regression analysis whereas R will give minimal output and store the results of a fit for subsequent interrogation or use with other R functions. This means that R can be tailored to produce exactly the analysis and results that you want rather than produce an analysis designed to fit all situations.

R is a language based product. This means that you interact with R by typing commands such as:

table(SEX, LIFE)

rather than by using menus, dialog boxes, selection lists, and buttons. This may seem to be a drawback but it means that the system is considerably more flexible than one that relies on menus, buttons, and boxes. It also means that every stage of your data management and analysis can be recorded and edited and re-run at a later date. It also provides an audit trail for quality control purposes.

R is available under UNIX (including Linux), the Macintosh operating system OS X, and Microsoft Windows. The method used for starting R will vary from system to system. On UNIX systems you may need to issue the R command in a terminal session or click on an icon or menu option if your system has a windowing system. On Macintosh systems R will be available as an application but can also be run in a terminal session. On Microsoft Windows systems there will usually be an icon on the Start menu or the desktop.

R is an open source system and is available under the GNU general public license (GPL) which means that it is available for free but that there are some restrictions on how you are allowed to distribute the system and how you may charge for bespoke data analysis solutions written using the R system. Details of the general public license are available from http://www.gnu.org/copyleft/gpl.html.

R is available for download from http://www.r-project.org/.

This is also the best place to get extension packages and documentation. You may also subscribe to the R mailing lists from this site. R is supported through mailing lists. The level of support is at least as good as for commercial packages. It is typical to have queries answered in a matter of a few hours.

Even though R is a free package it is more powerful than most commercial packages. Many of the modern procedures found in commercial packages were first developed and tested using R or S-Plus (the commercial equivalent of R).

When you start R it will issue a prompt when it expects user input. The default prompt is:

This is where you type commands that call functions that instruct R to (e.g.) read a data file, recode data, produce a table, or fit a regression. For example:

> table(SEX, LIFE)

If a command you type is not complete then the prompt will change to:

on subsequent lines until the command is complete:

> table(
+ SEX, LIFE +)

The > and + prompts are not shown in the example commands in the rest of this material.

The example commands in this material are often broken into shorter lines and indented for ease of understanding. The code still works as lines are split in places where R knows that a line is not complete. For example:

table(SEX,
      LIFE)

could be entered on a single line as:

table(SEX, LIFE)

In this example R knows that the command is not complete until the brackets are closed. The following example could also be written on one line:

salex.lreg.coeffs <-
  coef(summary(salex.lreg))

In this case R knows that the <- operator at the end of the first line needs further input.

R maintains a history of previous commands. These can be recalled and edited using the up and down arrow keys.

Output that has scrolled off the top of the output / command window can be recalled using the window or terminal scroll bars.

Output can be saved using the sink() function with a file name: sink("results.out") to start recording output. Use the sink() function without a file name to stop recording output: sink()

You can also use clipboard functions such as copy and paste to (e.g.) copy and then paste selected chunks of output into an editor or word processor running alongside R.

All the sample data files used in the exercises in this manual are space delimited text files using the general format:

R has facilities for working with files in different formats including (through the use of extension packages) ODBC (open database connectivity) and SQL data sources, EpiInfo, EpiData, Minitab, SPSS, SAS, S-Plus, and Stata format files.

Retrieving data

All of the exercises in this manual assume that the necessary data files are located in the current working directory. All of the data files that you require to follow this material are in a ZIP archive that can be downloaded from https://github.com/ernestguevarra/practical-r-for-epidemiologists/raw/master/data/prfe.zip.

A command such as:

read.table("data/fem.dat", header = TRUE)

retrieves the data stored in the file named fem.dat which is stored in the current working directory.

To retrieve data that is stored in files outside a different directory you need to specify the full path to the file. For example:

read.table("~/prfe/fem.dat", header = TRUE)

will retrieve the data stored in the file named fem.dat stored in the prfe directory under the user’s home directory on UNIX, Linux, and OS X systems.

R follows many UNIX operating and naming conventions including the use of the backslash (\) character to specify special characters in strings (e.g. using \n to specify a new line in printed output). Windows uses the backslash (\) character to separate directory and file names in paths. This means that Windows users need to escape any backslashes in file paths using an additional backslash character. For example:

read.table("c:\\prfe\\fem.dat", header = TRUE)

will retrieve the data that is stored in the file named fem.dat which is stored in the prfe directory off the root directory of the C: drive. The Windows version of R also allows you to specify UNIX-style path names (i.e. using the forward slash (/) character as a separator in file paths). For example:

read.table("c:/prfe/fem.dat", header = TRUE)

Path names may include shortcut characters such as:

`.`	The current working directory
`..`	Up one level in the directory tree
`~`	The user’s home directory (on UNIX-based systems)

R also allows you to retrieve files from any location that may be represented by a standard uniform resource locator (URL) string. For example:

read.table("file://~/prfe/fem.dat", header = TRUE)

will retrieve the data stored in the file named fem.dat stored in the prfe directory under the users home directory on UNIX-based systems.

All of the data files used in this section are stored in the /prfe directory Brixton Health’s website. This means, for example, that you can use the read.table() function specifying “https://github.com/ernestguevarra/practical-r-for-epidemiologists/blob/master/data/fem.dat” as the URL to retrieve the data that is stored in the file named fem.dat which is stored in the data/ directory of this guide’s GitHub repository.