Package 'Rcmdr' reference manual (2024)

Description

A platform-independent basic-statistics GUI (graphicaluser interface) for R, based on the tcltk package.

Translations

The R Commander comes with translations from English into several other languages. I am grateful to the following individuals and groups for preparing these translations:Basque, Jose Ramon Rueda;Brazilian Portuguese, Adriano Azevedo-Filho and Marilia Sa Carvalho;Catalan, Manel Salamero;Chinese, Tsungwu Ho, Frank C. S. Liu, and Cheng-shun Lee;Chinese (Simplified), Shulin Yang;French, Philippe Grosjean and Milan Bouchet-Valat;Galician, Anton Meixome;German: Friedrich Leisch and Gerhard Schoen;Greek: Vasileios Dimitropoulos, Anastasios Vikatos, and Andreas Vikatos;Indonesian, I Made Tirta;Italian, Stefano Calza;Japanese, Takaharu Araki;Korean, Chel Hee Lee, Dae-Heung Jang, and Shin Jong-Hwa;Polish, Lukasz Daniel;Romanian, Adrian Dusa;Russian, Alexey Shipunov;Slovenian, Jaro Lajovic and Matjaz Jeran;Spanish, Spanish R-UCA Project, http://knuth.uca.es/R.

Note: Many of these translations are not up-to-date. If you encounter an out-of-date translation and wish to take over updating and maintaining it, please contract the R Commander developers.

Author(s)

John Fox <[emailprotected]>, Manuel Munoz Marquez, and Milan Bouchet-Valat, with contributions from Liviu Andronic, Michael Ash, Theophilius Boye, Stefano Calza, Andy Chang, Vilmantas Gegzna,Philippe Grosjean, Richard Heiberger, G. Jay Kerns, Renaud Lancelot, Matthieu Lesnoff, Uwe Ligges, Samir Messad, Martin Maechler, Robert Muenchen, Duncan Murdoch, Erich Neuwirth, Dan Putler, Brian Ripley, Miroslav Ristic, Peter Wolf, and Kevin Wright.

Maintainer: John Fox <[emailprotected]>

References

Fox, J. (2017)Using the R Commander: A Point-and-Click Interface for R.Chapman and Hall/CRC Press.

Fox, J. (2005) The R Commander: A Basic Statistics Graphical User Interface to R. Journal of Statistical Software, 14(9): 1–42.

Description

Install Pandoc and LaTeX to increase the capabilities of the R Commander.

Details

The capabilities of the R Commander can be enhanced by installing additional software. The R Commander will work without this software but some features will not be activated.The following addiitional software can conveniently be installed via the R Commander Tools > Install auxiliary software menu. The resulting dialog box will takeyou to websites where the software can be downloaded. This menu item will only be displayedif one or more of these software packages are missing.

Here are the details:

• Pandoc: The Pandoc documentation-conversion software is used by the R Commander togenerate HTML (web), PDF, and Word files from the editable R Markdown document that is created by default during an R Commanderinteractive session. Pandoc is required by the rmarkdown package, which, along with the knitr package, performs these conversions.In the absence of Pandoc, R Markdown documents in the R Commander are processed by the older markdownpackage, which is capable only of producing HTML output. Pandoc is available from https://pandoc.org/installing.html.

  On Windows systems, Pandoc installs into a non-standard location in your user directory, typically C:\Users\<your user id>\AppData\Local\Pandoc\, and then places this subdirectoryon your system path. You may have to reboot for the change to your path to take effect, and Ihave found it necessary on two Windows 10 systems to re-run the Pandoc installer, first to uninstall Pandoc, and then to re-install it before it would be work.

• LaTeX: The LaTeX technical-typesetting system is required by the R Commander for PDF output from R Markdown or knitr documents produced during interactive R Commander sessions. In theabsence of LaTeX, direct PDF output is unavailable. Complete LaTeX systems are available for the variousplatforms that support R and the R Commander, including MikTeX from https://miktex.org/download for Windows systems; MacTeX from https://www.tug.org/mactex/ for MacOSX;and various sources (see https://www.latex-project.org/get/) for Linux/Unix systems.

See Also

Commander

Description

The CFA dialog is used to create and fit a confirmatory factor analysis model via thecfa and sem functions in the sem package.

Details

Select two or more variables for each factor by Control-clicking on their names in the variable-list box. Optionallygive the factor a name; this must be a valid R name. Then press the Define factor button.

Continue in this manner until all factors are specified. Note that if there are not at least two unique variables selected for each factor, themodel will probably be underidentified, causing sem to fail.

The radio buttons at the top of the dialog may be used to analyze either the correlation matrix or covariance matrix ofthe observed variables; to specify either correlated or orthogonal factors; and to identify the model either by settingthe factor variance to 1 or by setting the first loading for each factor to 1 (establishing a “reference indicator” forthe factor). A check box is provided to compute robust standard errors and tests.

Author(s)

John Fox [emailprotected]

See Also

sem, cfa

Description

Command simplifies the process of composing text-string versions of R commands, suitable, e.g., for doItAndPrint, from character strings representing a function name, arguments to the function, and (optionally) the name of a variable to be assigned the value returned by the function; all of the arguments to Command are text strings.

Usage

Command(fun, ..., to)Q(string)

Arguments

Value

Command returns the text representation of an R command, e.g., "model1 <- lm(y ~ x1 + x2, data=Dataset"; Q places quotation marks around string, e.g., Q("Annual Income"), and is meant to be used when an argument to Command should be quoted (see the last example below).

Author(s)

John Fox

See Also

doItAndPrint, justDoIt, logger.

Examples

Command("lm", "prestige ~ income + education", data="Duncan", weights="imp", to="model")Command("lm", "prestige ~ income + education", data="Duncan", weights=NULL, to="model")Command("plot", x="income", y="prestige", xlab=Q("Income"), ylab=Q("Prestige"))

Description

Start the R Commander GUI (graphical user interface)

Usage

Commander()

Details

Getting Started

For more detailed information about getting started, see Help -> Introduction to the R Commanderfrom the R Commander menus or Fox (2017).

The default R Commander interface consists of (from top to bottom) a menu bar, a toolbar,a code window with script and R Markdown tabs, an output window, and a messages window.

Commands to read, write, transform, and analyze data are entered using the menus in themenu bar at the top of the Commander window. Most menu items lead to dialog boxesrequesting further specification. I suggest that you explore the menus to seewhat is available.

Below the menu bar is a toolbar with (from left to right) an information fielddisplaying the name of the active data set; buttons for editing and displaying the activedata set; and an information field showing the active statistical model. There is alsoa Submit button for re-executing commands in the Script tab.The information fields for the active data set and active model are actually buttons that can be used to select the active data set and model from among, respectively,data frames or suitable model objects in memory.

Almost all commands require an active data set. When the Commander starts, there is no activedata set, as indicated in the data set information field. A data set becomes the active data setwhen it is read into memory from an R package or imported from a text file, SPSS data set,Minitab data set, STATA data set, SAS XPORT data set; or an Excel spreadsheet. In addition, the active data set can be selected from among R data framesresident in memory. You can therefore switch among data sets during a session.

By default, commands are logged to the Script tab (theinitially empty text window immediately below the toolbar), and commands and output appear in the Output window (the initially empty text window below theScript tab). Commands that don't require direct user interaction (such as interactiveidentification of points on a graph) are also used to create an R Markdown document inthe tab of the same name. When the R Markdown tab is in front, pressing the "Generate HTML report" buttoncompiles the document to create an html page with input and output, which opens in a web browser. To alter these andother defaults, see the information below on configuration. Note, for example, that the knitr packagecan be used to create a LaTeX document to be compiled to a PDF report, as an alternative to — or in addition to —an R Markdown document (see the use.knitr option below).

Some Rcmdr dialogs (those in the Statistics -> Fit models menu) produce linear, generalizedlinear, or other models. When a model is fit, it becomes the active model, as indicated in the informationfield in the R Commander toolbar.Items in the Models menu apply to the active model. Initially, there is no active model. If there are several models in memory, you can select the active model from among them.

If command logging in turned on, R commands that are generated from the menus and dialog boxes areentered into the Script and R Markdown tabs in the Commander. You can edit these commands in the normalmanner and can also type new commands. You can also type explanatory text in the R Markdown tab. Individual commandsin the Script tab can be continued over more than one line, but the several lines of a multi-line commandmust be submitted simultaneously. (It is not necessary, as in earlier versions of theR Commander, to begin continuation lines with white space.) The contents of the Script and R Markdowntabs can be saved during or at the end of the session, and a saved script or R Markdown document can be loaded into therespective tabs. The contents of the Output window can also be edited or saved to a text file. Finally,editing operations also work in the Messages window.

To re-execute a command or set of commands in the Script tab, select the lines to be executed usingthe mouse and press the Submit button at the right of the toolbar(or Control-R, for "run", or Control-Tab). If no text is selected, the Submit button (or Control-R or Control-Tab) submits the line containing the text-insertion cursor. Note that an error will be generated if the submittedcommand or commands are incomplete.

Pressing Control-F brings up a find-text dialog box (which can also be accessed via Edit -> Find) to search for text in the Script tab, R Markdown tab, knitr tab, Output window, or Messages window. Edit functions such assearch are performed in the Script tab unless you first click in another tab or window to make it active.

Pressing Control-S will save the Script tab, R Markdown tab, knitr tab, or Output window.

Pressing Control-A selects all of the text in the Script tab, R Markdown tab, knitr tab, Output window, or Messages window.

In addition, the following Control-key combinations work in these tabs and windows: Control-X, cut;Control-C, copy; Control-V, insert; Control-Z or Alt-Backspace, undo; and Control-W, redo.

Under Mac OS X, the command key may be used in place of the Control key, though thelatter works as well.

Right-clicking the mouse (clicking button 3 on a three-button mouse, or Control-left-clicking) in the tabs or windows brings upa "context" menu with the Edit-menu items, plus (in the Script, R Markdown, and knitr tabs) a Submit item.

You can open a larger editor window with the document in the Markdown or knitrtab by making the corresponding selection from the Edit menu, the right-clickcontext menu when the cursor is in the tab, or by pressing Control-E when thecursor is in the tab.

When youexecute commands from the Commander window, you must ensure that the sequence of commandsis logical. For example, it makes no sense to fit a statistical model to a data set that hasnot been read into memory.

Pressing a letter key (e.g., "a") in a list box will scroll the list box to bring the next entry starting with that letter to the top of the box.

You can cancel an R Commander dialog box by pressing the Esc key.

Most R Commander dialogs remember their state when this is appropriate, and can be restored to pristine state by pressing the Reset button.

Some R Commander dialogs have an Apply button that will execute the command generated by the dialog and then re-open the dialog in its previous state.

Exit from the Commander via the File -> Exit menu or by closing the Commander window.

Customization and Configuration

The preferred way of customizing the R Commander is to write a plug-in package:see help("Plugins").

Alternatively, configuration files reside in the etc subdirectory of the package, or in the locations given by the etc and etcMenus options (see below).

The Rcmdr menus can be customized by editing the file Rcmdr-menus.txt.

You can add R code to the package, e.g., for creating additional dialogs, by placing files withfile type .R in the etc directory, also editing Rcmdr-menus.txt to provideadditional menus, sub-menus, or menu-items. Alternatively, you can edit the source package and recompile it.

To reiterate, however, the preferred procedure is to write an R Commander plug-in package.

A number of functions are provided to assist in writing dialogs, and Rcmdr state informationis stored in a separate environment. See help("Rcmdr.Utilities")and the manual supplied in the doc directory of the Rcmdr package for more information.

In addition, several features are controlled by run-time options, set via the options("Rcmdr")command. These options should be set before the package is loaded. If the options are unset, which isthe usual situation, defaults are used. Specify options as a list of name=value pairs. You canset none, one, several, or all options. The available options are as follows:

ask.to.exit

if TRUE (the default), then the user is asked whether he or she wants to exit the Rcmdr;if this option is set to FALSE, then the subsequent option is also set to FALSE.

ask.on.exit

if TRUE (the default), then the user is asked whether to savethe script file, R Markdown file, and output file when the Rcmdr exits.

attach.data.set

if TRUE (the default is FALSE), the active data set is attached to thesearch path.

check.packages

if TRUE (the default), on start-up, the presence of all of the Rcmdrrecommended packages will be checked, and if any are absent, the Rcmdr will offer to install them.

command.text.color

Color for commands in the output window; the default is "red".

console.output

If TRUE, output is directed to the R Console, and the R Commander output window is not displayed. The default is FALSE, unless the R Commander is running under RStudio, in which case the default is TRUE.

crisp.dialogs

If TRUE, dialogs should appear on the screen fully drawn, rather than built upwidget by widget. Prior to R 2.6.1, this option only works on the Windows version of R, but should in any event be harmless.The default is TRUE. If you encounter stability problems, try setting this option toFALSE.

default.contrasts

Serves the same function as the general contrasts option; the default is
c("contr.Treatment", "contr.poly"). When the Commander exits, the contrastsoption is returned to its pre-existing value. Note that contr.Treatment is from the car package.

default.font.family

The default font for GUI elements such as menus and text labels, in the form of a Tk font family specification, given in a character string. For example, "Helvetica" specifies the sans-serif Helvetica font family.The default is taken from the TkDefaultFont. Normally a sans-serif font should be used.

default.font.size

The size, in points, of the default font. The default is 10 on non-Windowssystem and the size of the system font on Windows. To set the font size for R input and output, see the log.font.size option.The Rcmdr scale.factor option may also be used tocontrol font size.

discreteness.theshold

should be a positive integer; if greater than 0 (which is the default), the maximum number of distinct values for a numeric variable to be considered discrete; if 0 (or smaller), the threshold is taken as the smallest of 100,twice the squareroot of the number of cases in the active data set (n), and 10 times log10(n).

double.click

Set to TRUE if you want a double-click of the left mousebutton to press the default button in all dialogs. The default is FALSE.

editDataset.threshold

If the number of values in the currentdata set exceed this value (the default is 10000), then the standard Rdata editor is used in preference to the R Commander editDataset editor.

error.text.color

Color for error messages; the default is "red".

etc

Set to the path of the directory containing the Rcmdrconfiguration files; defaults to the etc subdirectory of the installed Rcmdr package.

grab.focus

Set to TRUE for the current Tk window to "grab" the focus — that is, to prevent the focus from being changed to another Tk window.On some systems, grabbing the focus in this manner apparently causes problems. The default is TRUE. If you experience focus problems, try setting this option to FALSE.

help_type

This Rcmdr option takes precedence over the global R help_typeoption (see options and help), and by default is set to "html".

iconify.commander

If TRUE, the Commander window is minimized on startup;the default is FALSE.

length.output.stack

The R Commander maintains a list of output objects, by default including the lastseveral outputs; the default length of the output stack is 10. popOutput() “pops” (i.e., returns and removes) the first entryof the output stack. Note that, as a stack, the queue is LIFO (“last in, first out”).

length.command.stack

The R Commander also maintains a list of commands that is managed similarly; the default lengthof this stack is also 10.

log.commands

If TRUE (the default), commands are echoed to the script window;if FALSE, the script window is not displayed.

log.font.family

The font family to be used for text in the script window, output window, messages window,etc., specified as a character vector giving a Tk font family. This should normally be a monospaced font like "Courier".The default is taken from the TkFixedFont.

log.font.size

The font size, in points, to be used in the script window,in the output window, messages window,in recode dialogs, and in compute expressions — that is, where a monospaced font is used. The default is 10.Alternatively the Rcmdr scale factor option may also be used tocontrol font size.

log.height

The height of the script window, in lines. The default is 10. Setting log.heightto 0 has the same effect as setting log.commands to FALSE.

log.text.color

Color for text in the script window; the default is "black".

log.width

The width of the script and output windows, in characters. The default is 80.

messages.height

The height of the messages window, in lines. The default is 4.

model.case.deletion

if TRUE (the default is FALSE), include a text box for case deletion in statistical-model dialog boxes (e.g., for linear models).

minimum.width

The minimum width, in pixels, for the main R Commander windows; the default is 1000.

minimum.height

The minimum height, in pixels, for the main R Commander windows; the default is 400.

multiple.select.mode

Affects the way multiple variables are selected in variable-list boxes.If set to "extended" (the default), left-clicking on a variable selects it and deselects any othervariables that are selected; Control-left-click togglesthe selection (and may be used to select additional variables); Shift-left-click extends the selection.This is the standard Windows convention. If setto "multiple", left-clicking toggles the selection of a variable and may be used to select more than onevariable. This is the behaviour in the Rcmdr prior to version 1.9-10.

number.messages

If TRUE, the default, messages in the messages window are numbered.

See Also

Statistik mit R leicht gemacht: Der R Commander – eine grafische Oberfläche | Statistik Dresden

open.graphics.devices

If TRUE (the default is FALSE), open the system graphics device and (if 3D RGL graphics are used) the RGL graphics device when the R Commander starts.

open.markdown.editor

If TRUE (the default is FALSE), open the R Markdown editor when the R Commander starts.

output.height

The height of the output window, in lines. The default is twice the heightof the script window, or 20 if the script window is suppressed. Setting output.heightto 0 has the same effect as setting console.output to TRUE.

output.text.color

Color for output in the output window; the default is "blue".

placement

Placement of the R Commander window, in pixels; the default is "",which lets the Tk window manager decide where to place the window; for example,"+20+20" should put the window near the upper-left corner of the screen, "-20+20" near theupper-right corner, though this doesn't appear to work reliably on Windows systems.

plugins

A character vector giving the names of Rcmdr plug-in packages to load when the Commander starts up. Plug-in packages can also be loaded from the Tools -> Load Rcmdr plug-in(s) menu.See Plugins.

prefixes

A four-item character vector to specify the prefixes used when output is directedto the R console; the default is c("Rcmdr> ", "Rcmdr+ ", "RcmdrMsg: ", "RcmdrMsg+ ").

quit.R.on.close

if TRUE, both the Commander and R are exited when the Commanderwindow is closed. The default is FALSE, in which case only the Commander is exited (and can be restartedby the command Commander()).

RcmdrEnv.on.path

If TRUE (the default is FALSE), the environment in which R Commander state informationis stored is placed on the search path. Some plug-ins, at least untilthey are updated, may require this setting.

retain.messages

If TRUE (the default), the contents of the message windoware not erased between messages. In any event, a "NOTE" message will not erase a preceding "WARNING" or "ERROR".

retain.selections

If TRUE (the default), dialogs remember their previous state, where appropriate,as long as the data set isn't changed; some dialogs, e.g., for probabilities, retain selections even when the data set chanages.

RExcelSupport

If TRUE (the default is FALSE), menus and output are handled by Excel.

rmarkdown.output

Values of several options for converting R Markdown to a document file. The default for this option is TRUE, which corresponds to markdown.output=list(command.sections=TRUE, section.level=3, toc=TRUE, toc_float=TRUE, toc_depth=3, number_sections=FALSE, translate.rmd.headers=TRUE). The sub-option command.sections controls whether most R commands produce sections in the R Markdown document; the sub-option section.level controls the level of the sections that are created; the sub-option translate.rmd.headers controls whether the headers are translated from English into another language, if a translation is available; and the other sub-options are standard for rmarkdown. The toc_float, toc_depth, and number_sections sub-options are only effective if Pandoc is installed.

rmd.output.format

The output file type for R Markdown documents if pandoc is installed; one of "html"(the default), "pdf" (requires LaTeX), "docx" (Word), or "rtf" (rich text file).

rmd.template

The quoted path to a .Rmd file to serve as a template forR code and output. The default is to use a template included with the package.

scale.factor

A scaling factor to be applied to all Tk elements, such as fonts. This works wellonly in Windows. The default is NULL.

scientific.notation

Higher numbers cause ordinary (decimal) notation to be increasingly preferred to scientific notationfor representing very small and very large numbers; correspond to the scipen option in R: see options. The defaultis 5, while the standard default in R is 0 (where 0 means that scientific notation is used whenever the resulting printedrepresentation of a number is smaller in scientific than in standard notation).

showData.threshold

a vector with 2 entries, defaulting to c(20000, 100). If the number of cases in the active data set exceeds the first number (default, 20,000)or the number of variables exceeds the second number (default, 100),then View() rather than showData() is used to display the data set. The reason for the option is that showData() is very slowwhen the number of cases or variables is large; setting the threshold to c(0, 0) suppresses the use of showData altogether. It's necessary to use showData however for the view of the active data set to be updated dynamically when, e.g., a variable is added.

show.edit.button

Set to TRUE (the default) if you want an Edit button in the Commanderwindow, permitting you to edit the active data set. Windows users may wish to set this option to FALSEto suppress the Edit button because changing variable names in the data editor can cause R to crash(though I believe that this problem as been solved).

sort.names

Set to TRUE (the default) if you want variable names to be sortedalphabetically in variable lists.

suppress.icon.images

Set to TRUE to suppress the icon images in dialog OK, Cancel, Reset, and Help buttons;the default is FALSE.

suppress.menus

if TRUE, the Commander menu bar and tool bar are suppressed, allowing anotherprogram (such as Excel) to take over these functions. The default (of course) is FALSE.

suppress.X11.warnings

On (some?) Linux and Mac OS X systems, multiple X11 warnings are generated by Rcmdrcommands after a graphics-device window has been opened. Set this option to TRUE (the default when runninginteractively under X11) to suppress reporting of these warnings. An undesirable side effectis that then all warnings and error messages are intercepted by the Rcmdr,even those for commands entered at the R command prompt. Messages produced by such commands will be printedin the Commander Messages window after the next Rcmdr-generated command.Some X11 warnings may be printed when you exit from the Commander.

theme

A ttk theme to control the overall style of the Commander GUI; should be one of the themes returned by tcltk2::tk2theme.list().The default theme varies by operating system, and can be discovered by entering the command tcltk2::tk2theme() in a fresh R session.

title.color

Color for the titles of some widgets, such as variable-list boxes; can be given as a color name, such as"blue" or as an RGB value, such as "#0000FF".The default is the standard color for ttk label frames, unless that is "#000000" or "black", in which case "blue" is used instead.

tkwait.commander

This option addresses a problem that, to my knowledge, is rare, and may occur on some non-Windows systems. If the Commander causes R to hang, then set the tkwait option toTRUE; otherwise set the option to FALSE or ignore it. An undesirable side effect of settingthe tkwait.commander option to TRUE is that the R session command prompt is suppressed until the Commander exits.One can still enter commands via the script window, however. In particular, there is no reason to use this optionunder Windows, and it should not be used with the Windows R GUI with buffered output when output is directed to theR console.

tkwait.dialog

If TRUE (the default is FALSE), R willwait until an R Commander dialog is closed. This has the disadvantage of preventinghelp pages from being displayed until a dialog is closed in the Mac OS X R.app and in RStudio. This was also the standard behavior of the R Commander in earlierversions and is provided for compatibility with previous behavior. If this option is TRUE,then the R Commander data editor is disabled in favor of the standard R platform-specificdata editor, and the new-data-set menu item is suppressed.

use.knitr

If TRUE (the default is FALSE), a knitr .Rnw LaTeX document is created in a tab of the main Commander window;this document can be compiled into .tex and .pdf reports via the knit2pdffunction in the knitr package.

use.markdown

If TRUE (the default is the negation of the use.knitr argument), an R Markdown documentis created, which can be compiled into an HTML, PDF, Word, or rich text file report.

use.rgl

If TRUE (the default), the rgl package will be loaded if it is present in anaccessible library; if FALSE, the rgl package will be ignored even if it is available. The rgl package can sometimes cause problems when running R under X11.

"valid.classes"

The classes of variables that the R Commander recognizes, in addition to numeric data; other variablesin a data set will be suppressed. The default is "factor", "ordered", "character", "logical", "POSIXct", "POSIXlt", "Date", "chron", "yearmon", "yearqtr", "zoo", "zooreg", "timeDate", "xts", "its", "ti", "jul", "timeSeries", "fts", "Period", "hms", "difftime").

variable.list.height

the number of items (typically variables) to display in list boxes; longer lists may be viewed by scrolling. The default is 6.

variable.list.width

a two-item vector controlling the width of list boxes, in characters, giving the minimum and maximumwidth to display; the default is c(20, Inf). If the widest item name falls in this range, then its number of charactersdetermines the width of the box. Note: This specification works only approximately.

warning.text.color

Color for warning messages; the default is "darkgreen".

Some options can also be set via the File -> Options menu, which will restart the Commander after options are set.

If you want always to launch the R Commander when R starts up, you can include the following code in one of R's start-up files (e.g., in the Rprofile.site file in R's etc subdirectory):

local({old <- getOption("defaultPackages")options(defaultPackages = c(old, "Rcmdr"))})

R Commander options can also be permanently set in the same manner.For more information about R initialization, see ?Startup.

Warning

The R Commander Script window does not provide a true console to R, and may have certain limitations.I don't recommend using the R Commander for serious programming or for data analysis that reliesprimarily on scripts — use a programming editor instead. If you encounter any problems with theScript tab, however, I'd appreciate it if you brought them to my attention.

Platform-Specific Issues

Under Windows, the Rcmdr package can be run under the Rgui in the SDI (single-documentinterface) mode, or under rterm.exe. You might experience problems running the Rcmdr underESS with NTEmacs or XEmacs, or under other R consoles. The R Commander can be run under the Rgui in MDI (multiple-document interface) mode but it is relatively inconvenient to do so and isn't recommended.

Occasionally, under Windows, after typing some text into a dialog box (e.g., a subsetting expression in the Subset Data Set dialog), buttons in the dialog (e.g., the OK button) will have no effect when they are pressed. Clicking anywhere inside or outside of the dialog box should restore the function of the buttons. As far as I have been able to ascertain, this is a problem with Tcl/Tk for Windows.I have not seen this behavior in some time and the problem may have been solved.

Under Mac OS X Mavericks and later, the R Commander may appear to freeze or hesitate when run under R.app if the R.app window is hiddenand "app nap" is turned on. It is recommended that app nap be turned off for R.app, which can be most conveniently done via theR Commander Tools menu. The app nap setting is permanent until changed and so the current setting will apply whether or not theR Commander is used. When R is first installed, app nap will be on for R.app. The tcltk package requires that X Windows isinstalled under Mac OS X, and as a consequence the Rcmdr package, which depends on tcltk, will not load if X Windows isabsent. X Windows for Mac OS X may be obtained from https://www.xquartz.org/.

Note

On startup, the R Commander sets options(na.action=na.exclude); this is done so thatobservation statistics such as residuals can be properly added to the active data setwhen there are missing values. Theoption is reset to its pre-existing value when the Commander exits. Some functions may notwork properly when the default na.action is set to na.exclude.

This version should be compatiable with the RExcel package, which can use the R Commandermenus.

Author(s)

John Fox [emailprotected]

References

Fox, J. (2017)Using the R Commander: A Point-and-Click Interface for R.Chapman and Hall/CRC Press.

Fox, J. (2005) The R Commander: A Basic Statistics Graphical User Interface to R. Journal of Statistical Software, 14(9): 1–42.

Fox, J. (2007)Extending the R Commander by "plug in" packages. R News, 7(3): 46–52.

See Also

Plugins, Rcmdr.Utilities, knit, knit2pdf

Examples

options(Rcmdr=list(log.font.size=12, default.contrasts=c("contr.Sum", "contr.poly")))

Description

Inicia la GUI (Interfaz Gráfica de Usuario) de R Commander

Usage

Commander()

Details

Empezando

La interfaz por defecto de R Commander consiste en (de arriba a abajo) una barra de menús, una barra de herramientas, una ventana de instrucciones, una ventana de salida y una ventana de mensajes.

Las instrucciones para leer, escribir, transformar y analizar datos se ejecutan usando la barra de menú de la parte superior de la ventana de R Commander. La mayor parte de los items de este menú le guiarán mediante ventanas de diálogo, preguntando más allá de la especificación. Es aconsejable explorar el menú para ver las opciones disponibles.

Bajo la barra de menú se encuentra la barra de herramientas con (de izquierda a derecha) un campo de información que muestra el nombre del conjunto de datos activos, botones para editar y mostrar el conjunto de datos activos y un campo de información mostrando el modelo estadístico activo. Bajo la ventana de instrucciones hay un botón Ejecutar para realizar las órdenes indicadas en la ventana de instrucciones. Los campos de información para los datos y el modelo activo son botones que pueden ser usados para seleccionar éstos entre, respectivamente, conjuntos de datos o modelos disponibles en memoria.

La mayor parte de las órdenes requiere un conjunto de datos activos. Cuando se ejecuta R Commander no hay conjunto de datos activos, como está indicado en el campo de información del conjunto de datos activos. Un conjunto de datos llega a ser un conjunto de datos activos cuando éste es leído en la memoria desde un paquete R o importado desde un archivo de texto, conjunto de datos SPSS, conjunto de datos Minitab, conjunto de datos STATA, Excel, Access o dBase. Además el conjunto de datos activos puede ser seleccionado desde conjuntos de datos R residentes en memoria. Los datos pueden ser elegidos de entre todos los conjuntos para cada sesión.

Por defecto, las órdenes son registradas en la ventana de instrucciones (la ventana de texto vacía inmediatamente después de la barra de herramientas); las órdenes y las salidas aparecen en la ventana de resultados (la ventana de texto vacía después de la ventana de instrucciones) y el conjunto de datos activos es adjuntado a la ruta de búsqueda. Para alterar éstos y otros parámetros por defecto, puede consultar la información pertinente en configuración.

Algunos diálogos de Rcmdr (éstos en Estadisticos -> Ajuste de modelos) generan el modelo lineal, modelo lineal generalizado y otros modelos. Cuando un modelo es ajustado, se convierte en el modelo activo, indicado en el campo de información de la barra de herramientas de R Commander. Los items del menú Modelos se aplican al modelo activo. Inicialmente, no hay modelo activo. Si hay varios modelos en memoria, puede elegir el modelo activo de entre ellos.

Si el registro de instrucciones está activo, las órdenes de R generadas desde los menús y los cuadros de diálogos, son introducidas en la ventana de instrucciones de R Comander. Se pueden editar estas órdenes de manera normal y se pueden escribir otras nuevas en la ventana de instrucciones. Las órdenes individuales pueden ser continuadas en más de una línea, pero cada línea después de la primera debe ser identada con uno o más espacios o tabuladores. El contenido de la ventana de instrucciones puede ser almacenado durante o al final de la sesión y un conjunto de instrucciones guardado puede ser cargado en la ventana de instrucciones. El contenido de la ventana de resultados puede ser editado o guardado en un archivo de texto.

Para volver a ejecutar una orden o un conjunto de ellas, se seleccionan las líneas que se desean ejecutar usando el ratón y se presiona el botón Ejecutar, situado a la derecha de la barra de herramientas (o Control-R, para ejecutarlos). Si no hay texto seleccionado el botón Ejecutar (o Control-R) envía el contenido de la línea que contiene el cursor de insercción. Observar que se generará un error si la orden o las órdenes enviadas son incompletas.

Presionando Control-F se abre un cuadro de diálogo de búsqueda de texto (también es accesible vía Editar -> Buscar) para buscar el texto en la ventana de instrucciones o la ventana de resultados. Las búsquedas son realizadas en la ventana de instrucciones a menos que primero pulse en la ventana de resultados para activarla.

Presionando Control-S se guardará el conjunto de instrucciones o la ventana de resultados.

Presionando Control-A se selecciona todo el texto del conjunto de instrucciones o de la ventana de resultados.

Pulsando el botón derecho del ratón (el tercer botón en un ratón de tres botones) en el conjunto de instrucciones o en la ventana de resultados se abre el menú contextual con los items del menú Editar, más un item Ejecutar (en la ventana de instrucciones).

Cuando ejecute órdenes en la ventana de R Commander, debe asegurarse que la sentencia sea lógica. Por ejemplo, no tiene sentido ajustar un modelo estadístico de un conjunto de datos que no ha sido leído en memoria.

Presionando una letra (ej. "a") en un cuadro con una lista se recorrerá ésta hasta la siguiente entrada que comience con esa letra desde el principio del cuadro.

Salir de R Commander se realiza mediante Fichero -> Salir o cerrando la ventana de R Commander.

Personalización y configuración

Los archivos de configuración están en el subdirectorio etc de cada paquete o en la localización dada por etc y en las opciones de etcMenus (mirar abajo).

Los menús de Rcmdr pueden ser personalizados editando el archivo Rcmdr-menus.txt.

Algunas funciones (ej. histograma) que normalmente no crean salida visible cuando se ejecutan desde la consola sí lo harán - a menos que se evite - cuando se ejecuten desde la ventana de instrucciones de R Commander. Tal salida puede ser suprimida listando los nombres de estas funciones en el archivo log-exceptions.txt.

Puede añadir código R al paquete, ej., para crear diálogos adicionales, colocando archivos con extensión .R en el directorio etc, además puede editar Rcmdr-menus.txt para proporcionar menús adicionales, submenús o items. Una demostración de esto se proporciona mediante el archivo BoxCox.demo. Para activar la demo, renombre el archivo a BoxCox.R y descomente la correspondiente línea del menú en Rcmdr-menus.txt. De forma alternativa, puede editar el código del paquete y recompilarlo.

Algunas funciones son proporcionadas para ayudar a escribir diálogos y la información del estado de Rcmdr en un emplazamiento separado. Mirar help("Rcmdr.Utilities") y el manual suministrado en el directorio doc del paquete de Rcmdr para mayor información.

Además, varias características son controladas mediante opciones, en tiempos de ejecución, establecidas por la orden options("Rcmdr"). Estas opciones deben ser establecidas antes de cargar el paquete. Si las opciones no están establecidas, que es la situación normal, serán usados los parámetros por defecto. Las opciones se especifican como una lista de pares name$=$values. Puede no establecer, establecer una, varias, o todas las opciones. Las opciones disponibles son las dadas a continuación:

attach.data.set

Si es TRUE (por defectoFALSE), el conjunto de datos activo es fijado como la ruta debúsqueda.

check.packages

Si es TRUE (por defecto), alarranque, la presencia de todos los paquetes recomendados de Rcmdrserán comprobados y si alguno no está instalado, Rcmdr preguntará si deben instalarse.

command.text.color

El color de las órdenes en la ventana de resultados es, por defecto, "red".

console.output

Si es TRUE la salida será dirigida ala consola de R y la ventana de salida de R Commander no semostrará. Por defecto es FALSE.

contrasts

Ofrece la misma función que la opción generalcontrasts; por defecto es c("contr.Treatment", "contr.poly"). Cuando se sale de Commander la opcióncontrasts vuelve a su valor preexistente. Observe quecontr.Treatment es del paquete car.

crisp.dialogs

Si es TRUE, los diálogos debenaparecer en la pantalla dibujada completamente, más que acumulardispositivo a dispositivo. Esta opción debería afectar sólo aversiones Windows de R, pero debe en cualquier caso serinofensivo. Por defecto es TRUE bajo versiones Windows de R2.1.1 y superiores y FALSE si no. Si está trabajando enWindows y encuentra que se incrementan los problemas de estabilidad,pruebe establecer esta opción a FALSE.

default.font

La fuente por defecto, como la especificaciónde la fuente de X11, dada en cadena de caracteres. Si estáespecificado, este valor toma precedencia sobre el tamaño de la fuentepor defecto (abajo). Esta opción es sólo para sistemas no-Windows.

default.font.size

Tamaño, en puntos, por defecto de lafuente. Por defecto es 10 para sistemas Windows y 12 para otrosistemas, salvo especificación de lo contrario (mirar el itemanterior). La fuente por defecto es"*helvetica-medium-r-normal-*-xx*", donde xx es pordefecto el tamaño de la fuente. Esta opción es sólo para sistemas no-Windows.

double.click

Establecer a TRUE si quiere que un doble click con el botón izquierdo del ratón sirva para pulsar el botón por defecto en todos los diálogos. Por defecto es FALSE.

error.text.color

Color de los mensajes de error; pordefecto es "red".

etc

Establece la ruta del directorio que contiene los archivos de configuración de Rcmdr; por defecto el subdirectorio etc del paquete Rcmdr instalado.

grab.focus

Establecer a TRUE para "mantener" el enfoque en la ventana actual de Tk, esto es, para prevenir que el enfoque sea cambiado a otra ventana Tk. En algunos sistemas, mantener el enfoque de esta forma, puede causar problemas. Por defecto es TRUE. Si experimenta problemas de enfoque, intente establecer esta opción a FALSE.

load.at.startup

Vector de caracteres de nombres de los paquetes que deben ser cargados cuando el paquete Rcmdr es cargado; por defecto se carga sólo el paquete car. Otros paquetes requeridos serán cargados cuando se necesiten. Si esto está disponible, el paquete car será cargado cuando Commander se inicie en cualquier caso.

log.commands

Si es TRUE (por defecto), los comandos son repetidos en la ventana de instrucciones; si es FALSE, la ventana de instrucciones no se muestra.

log.font.size

Tamaño de la fuente, en puntos, que es usado en la ventana de instrucciones, en la ventana de resultados, en diálogos recodificados y en expresiones de cáculo, esto es, donde es usada una fuente monoespacio. Por defecto es 10 para sistemas Windows y 12 para otros sistemas.

log.height

La altura de la ventana de instrucciones, en líneas. Por defecto es 10. Estableciendo log.height a 0 tiene el mismo efecto que establer log.commands a FALSE.

log.text.color

Color del texto de la ventana de instrucciones; por defecto es "black".

log.width

La anchura de la ventana de instrucciones y la de salida, en caracteres. Por defecto es 80.

multiple.select.mode

Afecta a la forma en la que múltiples variables son seleccionadas en una caja de listas de variables. Si se establece a "extended" (por defecto), el botón izquierdo en una variable selecciona ésta y deselecciona cualquier otra variable que estuviera seleccionada; Control+botón izquierdo acciona la selección (y puede ser usado para seleccionar variables adicionales); Mayúsculas+botón izquierdo extiende la selección. éste es el convenio estándar de Windows. Si lo establece a "multiple", el botón izquierdo acciona la selección de una variable y puede ser usado para seleccionar más de una variable. éste es el comportamiento de Rcmdr antes de la versión 1.9-10.

output.height

Altura de la ventana de resultados, en líneas. Por defecto es dos veces la altura de la ventana de instrucciones o 20 si la ventana de instrucciones es suprimida. Establecer output.height a 0 tiene el mismo efecto que console.output a TRUE.

output.text.color

Color de la salida en la ventana de resultados, por defecto es "blue".

placement

Emplazamiento de la ventana de R Commander, en píxeles; por defecto es "$-40+20$", lo que pone la ventana cerca de la esquina superior derecha de la pantalla.

plugins

Vector de caracteres con los nombres de paquetes de plugins de Rcmdr a cargar cuando Commander arranque. Los paquetes plugins también pueden ser cargados desde el menú Herramientas -> Cargar paquete(s).

suppress.menus

Si es TRUE, la barra de menús y de herramientas de R Commander son suprimidas, permitiendo que otro programa (como Excel) asuma esas funciones. Por defecto (por supuesto) es FALSE.

suppress.X11.warnings

En (algunos) sistemas Linux X11 se generan múltiples advertencias por las órdenes de Rcmdr, después de abrir la ventana del dispositivo gráfico. Establecer esta opción a TRUE (por defecto cuando arranca interactivamente bajo X11 antes de la versión de R 2.4.0) suprime la aparición de estas advertencias. Un efecto secundario indeseable es que entonces todas las advertencias y mensajes de error son interceptados por Rcmdr, incluso para las instrucciones introducidas en los avisos de R. Los mensajes producidos por tales órdenes serán impresos en la ventana de mensajes de R Commander después de la siguiente orden generada en Rcmdr. Algunas advertencias de X11 puede ser impresas al salir de R Commander. Este problema sólo se aplica a versiones de R anteriores a 2.4.0 y el valor por defecto de la opción es establecido por consiguiente.

retain.messages

Si es TRUE (por defecto FALSE), el contenido de la ventana de mensajes no es borrado entre mensajes. En cualquier caso, un mensaje "NOTE" no borrará un anterior "WARNING" o "ERROR".

RExcelSupport

Establecido como TRUE (por defecto es FALSE), los menús y salidas son dirigidas a Excel.

scale.factor

Factor de escala aplicado a todos los elementos Tk, como las fuentes. Esto funciona bien sólo en Windows. Por defecto es NULL.

showData.threshold

Si el número de variables en el conjunto de datos activos excede este valor (por defecto, 100), entonces edit(), más que showData(), es utilizado para exhibir el conjunto de datos. Un inconveniente es que el control no se devuelve a Commander hasta que la ventana de edición sea cerrada. La razón de esta opción es que showData() es muy lento cuando el número de variables es grande; fijando el umbral a 0 suprime el uso en conjunto de showData.

show.edit.button

Fijar a TRUE (por defecto) si quiere un botón Editar en la ventana de Commander, que permita editar el conjunto activo de datos. Los usuarios de Windows pueden desear establecer esta opción a FALSE para suprimir el botón Editar porque cambiando los nombres de las variables en el editor de datos se puede causar que R falle (aunque este problema se cree solucionado).

sort.names

Fijar a TRUE (por defecto) si se quiere ordenar alfabéticamente el nombre de las variables en una lista de variables.

tkwait

Esta opción trata un problema que, en miconocimiento, es raro y puede ocurrir en algunos sistemas noWindows. Si R Commander causa que se cuelgue R, entonces establezca laopción tkwait a TRUE; o conserve la opción enFALSE e ignórelo. Un indeseable efecto secundario de establecerla opción tkwait a TRUE es que el aviso de órdenes de lasesión de R es suprimido hasta salir de R Commander. Uno sin embargo todavía puede introducir órdenes por la ventana de instrucciones. Enparticular, no hay razón para usar esta opción bajo Windows y no sedebería usar con la GUI de R en Windows con salida protegida cuando lasalida esté dirigida ala consola de R.

use.rgl

Si es TRUE (por defecto), el paquete rgl será cargado si está presente en una librería accesible, si es FALSE, el paquete rgl será ignorado aunque esté disponible. El paquete rgl puede a veces causar problemas cuando se arranca R bajo X11.

warning.text.color

Color de los mensajes de advertencia;por defecto es "darkgreen".

Muchas opciones pueden también ser establecidas mediante el menú Archivo -> Opciones, que reiniciará R Commander después de que las opciones sean establecidas.

Si quiere lanzar R Commander cuando inicie R, puede incluir la siguienteinstrucción en uno de los ficheros de inicio de R (por ejemplo, en elfichero Rprofile.site de la carpeta etc de R):

local({old <- getOption("defaultPackages")options(defaultPackages = c(old, "Rcmdr"))})

Las opciones de R Commander puede ser establecidas de forma permanentede la misma forma. Para más información sobre el inicio de R, véase ?Startup.

Avisos

La ventana de instrucciones de R Commander no proporciona una verdadera consola a R y tiene ciertas limitaciones. No se recomienda usar R Commander para la programación sería o el análisis de datos que confíe primordialmente en instrucciones - usar un editor de programación en su lugar. Por ejemplo, para declaraciones de composiciones de R incluidas entre llaves "\{ \}", incluyendo definición de funciones, no serían analizadas ni ejecutadas correctamente, aunque si las líneas después de las primeras que estén identandas. Puede ejecutar declaraciones de composiciones desde la ventana de instrucciones separando los comandos dentro de las llaves por puntos y comas.

Problemas Conocidos

Ocasionalmente, bajo Windows, después de teclear algún texto en un cuadro de diálogo (ej. subconjunto de expresiones en el diálogo de subconjunto de conjunto de datos), algunos botones en el diálogo (ej. el botón Aceptar) pueden no tener efecto cuando sean presionados. Pulsando en cualquier sitio, dentro o fuera del cuadro de diálogo, debería restaurarse las funciones de los botones. Por lo que se ha podido comprobar, éste es un problema con Tcl/Tk de Windows.

Note

Bajo Windows, el paquete Rcmdr puede también funcionar bajo de Rgui en modo SDI (interfaz de único documento) o bajo rterm.exe; puede ser que experimente problemas ejecutando Rcmdr bajo ESS con NTEmacs o XEmacs.

Author(s)

John Fox [emailprotected] (de la versión inglesa)

Manuel González (traductor) [emailprotected]

Manuel Muñoz Márquez (traductor–revisor) [emailprotected]

Véase http://knuth.uca.es/R/doku.php?id=equipotraduccion

See Also

Plugins

Examples

options(Rcmdr=list(log.font.size=12, contrasts=c("contr.Sum", "contr.poly")))

Description

The compute dialog is used to compute new variables.

Details

The name of the new variable must bea valid R object name (consisting only of upper and lower-case letters, numerals, andperiods, and not starting with a numeral).

Enter an R expression in the box at the right. The expression is evaluated using the active data set. You can double-click in the variable-list box to enter variable names in the expression. The expressionmust evaluate to a valid variable, which is added to the active data set.

Author(s)

John Fox [emailprotected]

See Also

Arithmetic

Description

Allows the user to enter a new dataset, modify data values in anexisting dataset, add rows or columns to the dataset, or delete rows or columns.

Usage

editDataset(data, dsname, ...)## S3 method for class 'data.frame'editDataset(data, dsname, ...)## S3 method for class 'character'editDataset(data, dsname, ...)## S3 method for class 'NULL'editDataset(data, dsname, ...)

Arguments

Details

editDataset is a straightforward spreadsheet-like data editor, suitable for editing data frames that are not too large (say smaller than about 10,000 values). It is defined as a generic function with a data.frame method to allow for objects with unique properties that inherit from the data.frame class. The character and NULL methods permit editing an initially empty data set.

• Use the mouse and the arrow keys to navigate the cells of the data table,including the row and column names.

• Columns consisting only of numbers will produce numeric variables inthe data frame constructed by editDataset; columns with any non-numericvalues will produce factors or (if they contain only the valuesTRUE and FALSE) logical variables.

• When entering values with embedded blanks, it is permissible but not necessary to enclose the values in quotes(e.g, "some PS" or 'less than HS').

• Clicking in a cell and typing a new value replaces the previous value.

• Row and column names can be modified in the same manner.

• Double-clicking in a cell deletes the previous value and replaces it with NA.

• Enlarge the data set by pressing the Add row or Add column button at the top of the data editor; the new row or column will initially be filled with NAs and will have an auto-generated row or column name. Pressing the Enteror Return key will also add a row; pressing the Tab key will also adda column.

• Right-clicking (or Control-clicking, or under Mac OS X Command-clicking) brings up a context menu, permitting several operationson cells, rows, and columns, including deleting the current row or column.

• Similarly, several actions are available via the Edit menu.

• The key-combinations Control-x, and Control-v, may also be usedrespectively to cut, copy, and paste cell values. (Under Mac OS X, Command-x, Command-c, andCommand-v also work.)

• Pressing the OK button or selecting Exit and save from theFile menu exits the data edtior and saves the edited data set to the globalenvironment. Pressing the Cancel button or selecting Cancel from theFile menu exits the editor discarding the edited data set.

Value

This function does not return a useful value, but has the sideeffect of modifying or creating a data set in the global environment.

Note

editDataset is limited to editing data frames that are composed only ofnumeric, factor, and logical columns.

Author(s)

John Fox [emailprotected]

See Also

edit.data.frame, for the standard R data editor.

Examples

if (interactive()) editDataset()

Description

This dialog is used to specify a generalized linear model to be fit by the glm function.

Details

The left model-formula box specifies the response variable to be used in the model; itmay be a variable name or an expression evaluating to the response variable, such asworking == "Fulltime".

The right model-formula box specifies the right-hand (i.e., predictor) side of the model.See glm for details.

You can type directly in the model formula boxes. Alternatively, double-clicking the left mousebutton on a variable in the variable-list transfers it to the left-hand side of the model (ifit is empty or selected) or to the right-hand side. Factors are indicated in the variable list; all other variables are numeric.You can also enter operators and parentheses using the buttonsabove the formula. If you select several variables in the variable-list box, clicking on the+, *, or : button will enter them into the model formula.

Double-click the left mouse button to select a family in the "Family" box and the correspondingpermissible link functions appear in the "Link function" box to the right. Initially, thecanonical link for the family is selected. See family for details.

Specifying a subset expression allows you to fit the model to a subset of observationsin the active data set. For example, assuming that gender is a variable in theactive data set, entering gender == "Male" would restrict the model to males.

The weights box allows you to select a variable specifying prior weights from thedrop-down list. Weights giving numbers of trials may be used, for example, to fit abinomial GLM; in this case, the response variable should give the proportion of"successes" for each binomial observation. Click in the weights combo box to see a list of numeric variables in thecurrent data set; type a letter in the box to move the selection cursor to the nextvariable beginning with that letter.

There is an optional case-deletion box, whose presence is controlled by the model.case.deletion R Commander option (see Commander). Typing the row numbers (e,g., 6 16) or row names (e.g., minister conductor) of cases to be deleted removes these cases from the fitted linear model. Row names with embedded blanks must be quoted (e.g., "railroad engineer"), in which case all row names specified should be quoted (e.g., "railroad engineer" "minister"). You cannot specify both a subset expression and case deletion.

If the active model is a generalized linear model, and the active data set has not changed, then the initial values of the left-hand-side,right-hand-side, family, link, weights, and subset fields are retained from the active model.

Author(s)

John Fox [emailprotected]

See Also

glm, family, Comparison

Description

This dialog is used to specify a hierarchical cluster analysis solution using hclust, with the distance matrix calculated usingdist.

Details

Enter a name for the hierarchical clustering solution to be created if you want to retain more than one solution. The solution name must be a valid R object name (consisting only of upper- and lower-case letters, numerials, and periods, and not starting with a number).

Select the variables to be included in the solution using the variable selection box on the left side of the dialog box. A non-contiguous set of variables can be selected by pressing your control key (ctrl) while selecting variables.

Specifying a subset expression (the field below the variable selection box) allows you to obtain a clustering solution for a subset of observations in the active data set. For example, assuming that gener is a variable in the active data set, entering gender == "Male" would restrict the solution to males.

Select a clustering method and a distance measure if you are working with raw data. There is often a relationship between the selection of these two items. For example, squared-euclidian distance is appropriate for Ward's method of cluster analysis. If your data is a distance matrix, then select "No Transformation" as the distance measure.

The "Plot Dendrogram" option results in the dendrogram of the solution being display by using the plot function.

Author(s)

Dan Putler

See Also

hclust, dist

Description

This dialog is used to specify a linear model to be fit by the lm function.

Details

The left model-formula box specifies the response variable to be used in the model; itmay be a variable name or an expression evaluating to the response variable, such aslog(income).

The right model-formula box specifies the right-hand (i.e., predictor) side of the model.See lm for details.

You can type directly in the model formula boxes. Alternatively, double-clicking the left mousebutton on a variable in the variable-list transfers it to the left-hand side of the model (ifit is empty or selected) or to the right-hand side. You can also enter operators and parentheses using the buttonsabove the formula. If you select several variables in the variable-list box, clicking on the+, *, or : button will enter them into the model formula.

Specifying a subset expression allows you to fit the model to a subset of observationsin the active data set. For example, assuming that gender is a variable in theactive data set, entering gender == "Male" would restrict the model to males.

The weights box allows you to perform weight-least-squares (WLS) regression;select a weight variable from the drop-down list. Click in the weights combo box to see a list of numeric variables in thecurrent data set; type a letter in the box to move the selection cursor to the nextvariable beginning with that letter.

There is an optional case-deletion box, whose presence is controlled by the model.case.deletion R Commander option (see Commander). Typing the row numbers (e,g., 6 16) or row names (e.g., minister conductor) of cases to be deleted removes these cases from the fitted linear model. Row names with embedded blanks must be quoted (e.g., "railroad engineer"), in which case all row names specified should be quoted (e.g., "railroad engineer" "minister"). You cannot specify both a subset expression and case deletion.

If the active model is a linear model and the active data set has not changed, then the initial values of the left-hand-side,right-hand-side, weights, and subset fields are retained from the previous model.

Author(s)

John Fox [emailprotected]

See Also

lm, Comparison

Description

Plug-ins are R packages that extend the R Command interface.

Details

An R Commander plug-in is an ordinary R package that (1) provides extensions to the R Commander menus is a file named menus.txt located in the package's etc directory; (2) provides call-back functions required by these menus;and (3) in an RcmdrModels: field in thepackage's DESCRIPTION file, augments the list of model objects recognized by the R Commander. The menus provided by a plug-in package aremerged with the standard Commander menus. It is also possible to remove menusand menu items from the standard Commander menu file or from the files ofplug-ins installed before the current one.

Plug-in packages given in the R Commander plugins option (see Commander) are automatically loaded when the Commanderstarts up. Plug-in packages may also be loaded via the CommanderTools -> Load Rcmdr plug-in(s) menu; a restart of the Commander isrequired to install the new menus. Finally, loading a plug-in package whenthe Rcmdr is not loaded will load the Rcmdr and activate theplug-in.

An illustrative R Commander plug-in package, RcmdrPlugin.TeachingDemos, isavailable on CRAN.

A variety of utility functions is available to support R Commander plug-in packages;see Rcmdr.Utilities.

For more details, see Fox, Writing R Commander Plug-in Packages at https://socialsciences.mcmaster.ca/jfox/Books/RCommander/Writing-Rcmdr-Plugins.pdf.

See Also

Commander, Rcmdr.Utilities

Description

For technical reasons involving Rcmdr plug-in packages, the Rcmdr packagere-exports a number of functions from the tcltk package that are used by Rcmdrutilities.

It also re-exports the showdata function from the relimp package, which is uses internally to display data sets.

Author(s)

John Fox [emailprotected]

See Also

tcltk

Description

These functions support writing additions to the Rcmdr package, preferably by writingan Rcmdr plug-in package. Although it is not recommended, additional R codecan also be placed in files with file type .R in the etc subdirectoryof the Rcmdr package. In this case, you can add menus, submenus, and menu items by editing the fileRcmdr-menus.txt in the same directory.

Usage

activateMenus()activeDataSet(dsname, flushModel=TRUE, flushDialogMemory=TRUE)ActiveDataSet(name)activeDataSetP()activeModel(model)ActiveModel(name)activeModelP()anovaP()beginRmdBlock()beginRnwBlock()Character(names)characterP(n=1)checkActiveDataSet()checkActiveModel()checkBoxes(window=top, frame=stop("frame not supplied"), boxes=stop("boxes not supplied"), initialValues=NULL, labels=stop("labels not supplied"), title=NULL, ttk=FALSE, columns=1) # macrocheckClass(object, class, message=NULL) # macrocheckFactors(n=1)checkMethod(generic, object, message=NULL, default=FALSE, strict=FALSE, reportError=TRUE) # macrocheckNumeric(n=1)checkReplace(name, type=gettextRcmdr("Variable"))checkTwoLevelFactors(n=1)checkVariables(n=1)closeCommander(ask=TRUE, ask.save=ask)closeDialog(window, release=TRUE) # macroCoef(object, ...)Predictors(type=c("all", "numeric", "factor"))PredictorsP(n=1, type=c("all", "numeric", "factor"))CommanderWindow()dataSetsP(n=1)defmacro(..., expr)dialogSuffix(window=top, onOK=onOK, onCancel=onCancel, rows, columns,focus=top, bindReturn=TRUE, preventGrabFocus=FALSE, preventDoubleClick=FALSE, preventCrisp, use.tabs=FALSE, notebook=notebook, tabs=c("dataTab", "optionsTab"), tab.names=c("Data", "Options"), grid.buttons=FALSE, resizable=FALSE, force.wait=FALSE) # macroDiscreteNumeric(names)discreteNumericP(n=1)doItAndPrint(command, log=TRUE, rmd=log)EffectP()endRmdBlock()endRnwBlock()enterMarkdown(command)enterKnitr(command)errorCondition(window=top, recall=NULL, message, model=FALSE) # macroexists.method(generic, object, default=TRUE, strict=FALSE)Factors(names)factorsP(n=1)formulaFields(model, hasLhs=TRUE, glm=FALSE)flushDialogMemory(what)gassign(x, value)getCases(cases, remove=TRUE)getDialog(dialog, defaults=NULL)## S3 method for class 'combobox'getFrame(object)## S3 method for class 'listbox'getFrame(object)## S3 method for class 'combobox'getSelection(object)## S3 method for class 'listbox'getSelection(object)getRcmdr(x, mode="any", fail=TRUE)gettextRcmdr(...)glmP()GrabFocus(value)groupsBox(recall=NULL, label=gettextRcmdr("Plot by:"), initialLabel=gettextRcmdr("Plot by groups"), errorText=gettextRcmdr("There are no factors in the active data set."), variables=Factors(), plotLinesByGroup=FALSE, positionLegend=FALSE, plotLinesByGroupsText=gettextRcmdr("Plot lines by group"), initialGroup=NULL, initialLinesByGroup=1, window=top) # macrogroupsLabel(frame=top, groupsBox=groupsBox, columnspan=1, initialText=NULL, ratio=FALSE) # macrohclustSolutionsP()initializeDialog(window=top, title="", offset=10, preventCrisp, use.tabs=FALSE, notebook=notebook, tabs=c("dataTab", "optionsTab"), suppress.window.resize.buttons=TRUE) # macroinsertRmdSection(text)is.valid.name(x)is.valid.number(string)is.SciViews()justDoIt(command)knitrP()Library(package, pos=length(search()), rmd=TRUE)listAllModels(envir=.GlobalEnv, ...)listAOVModels(envir=.GlobalEnv, ...)listCharacter(dataSet=ActiveDataSet())listDataSets(envir=.GlobalEnv, ...)listDiscreteNumeric(dataSet=ActiveDataSet())listFactors(dataSet=ActiveDataSet())listGeneralizedLinearModels(envir=.GlobalEnv, ...)listLinearModels(envir=.GlobalEnv, ...)listMultinomialLogitModels(envir=.GlobalEnv, ...)listNumeric(dataSet=ActiveDataSet())listPlugins(loaded=FALSE)listProportionalOddsModels(envir=.GlobalEnv, ...)listTwoLevelFactors(dataSet=ActiveDataSet())listVariables(dataSet=ActiveDataSet())lmP()logger(command, rmd=TRUE)logLikP()LogWindow()MacOSXP(release)manualTranslationP()MarkdownP()mavericksP()Message(message, type=c("note", "error", "warning"))MessagesWindow()modelCapability(capability)modelFormula(frame=top, hasLhs=TRUE, rhsExtras=NULL, formulaLabel=gettextRcmdr("Model Formula"), showBar=FALSE) # macromodelsP(n=1)multinomP()Numeric(names)numericP(n=1)OKCancelHelp(window=top, helpSubject=NULL, model=FALSE, reset=NULL, apply=NULL, helpPackage=NULL) # macroOutputWindow()packageAvailable(name)polrP()popCommand(keep=FALSE)popOutput(keep=FALSE)putDialog(dialog, values=NULL, resettable=TRUE)putRcmdr(x, value)radioButtons(window=top, name=stop("name not supplied"), buttons=stop("buttons not supplied"), values=NULL, initialValue=..values[1], labels=stop("labels not supplied"), title="", title.color=getRcmdr("title.color"), right.buttons=FALSE, command=function(){}, columns=1) # macroRappP()RcmdrEditor(buffer, title="R Commander Editor", ok, help=NULL, file.menu=NULL, edit.menu=NULL, context.menu=NULL, toolbar.buttons=NULL)RcmdrTclSet(name, value)RcmdrTkmessageBox(message, icon=c("info", "question", "warning", "error"), type=c("okcancel", "yesno", "ok"), default, title="")removeLastRmdBlock()removeLastRnwBlock()removeNullRmdBlocks()removeNullRnwBlocks()removeStrayRmdBlocks()removeStrayRnwBlocks()RExcelSupported()rglLoaded()RmdWindow()RnwWindow()setBusyCursor()setIdleCursor()sortVarNames(x)subOKCancelHelp(window=subdialog, helpSubject=NULL) # macrosubsetBox(window = top, subset.expression = NULL, model = FALSE) # macrosuppressMarkdown(command)tclvalue(x)titleLabel(...)tkfocus(...)tkspinbox(parent, ...)trim.blanks(text)TwoLevelFactors(names)twoLevelFactorsP(n=1)UpdateModelNumber(increment=1)variableComboBox(parentWindow, variableList=Variables(), export="FALSE", state="readonly", initialSelection=gettextRcmdr(nullSelection), title="", nullSelection="<no variable selected>", adjustWidth = FALSE)variableListBox(parentWindow, variableList=Variables(), bg="white", selectmode="single", export="FALSE", initialSelection=NULL, listHeight=getRcmdr("variable.list.height"), title)Variables(names)varPosn(variables, type=c("all", "factor", "numeric", "nonfactor", "twoLevelFactor"), vars=NULL)WindowsP()X11P()# the following function is exported for technical reasons,# but are not meant to be called directlycommanderPosition()

Arguments

Details

There are several groups of functions exported by the Rcmdr package and documented briefly here.To see how these functions work, it is simplest to examine the dialog-generating functions in the Rcmdr package.Also see the RcmdrPlugin.survival package for examples.

Executing and logging commands:

For conveniently constructing character strings representing commands, see Command.

The functions doItAndPrint, justDoIt, and logger control theexecution, logging, and printing of commands generated by menus and dialogs.logger(command) adds command to the log/script window and to the output window.justDoIt(command) causes commandto be executed. doItAndPrint(command) does both of these operations, and also prints theoutput produced by the command.The R Commander maintains a list of output objects, by default including the last10 outputs. popOutput() “pops” (i.e., returns and removes) the first entryof the output stack. Note that, as a stack, the queue is LIFO (“last in, first out”). Use popOutput(keep=TRUE) to access the last output but keep it in the stack.There is also a stack of commands, which is accessed similarly by popCommand().Ocassionally, it's necessary to assign an object directly in the global environment, andthis can be done with the gassign function.

Normally commands also generate an R Markdown block.suppressMarkdown takes a command in character-string form and adds an attribute to itthat will cause the command not to be entered in the R Markdown tab. This is useful when acommand, such as identify, requires direct user interaction and won't generate useful Markdown.enterMarkdown can be used to enter command blocks directly in the R Markdown tab; this shouldrarely be required. The functions beginRmdBlock, endRmdBlock, removeNullRmdBlocks,removeLastRmdBlock, and removeStrayRmdBlocks should normally not be called directly.The functions enterKnitr,beginRnwBlock, endRnwBlock, removeNullRnwBlocks, removeLastRnwBlock,and removeStrayRnwBlocks perform similar functions for Knitr documents.

insertRmdSection Inserts a Markdown section title immediately above the last R command block, with the specified text as the title. In most instances it's unnecessary to do this directly because most commands automatically generate a Markdown section title.

Checking for errors:

The function is.valid.name checks whether a character string specifies a valid namefor an R object. The function is.valid.number checks whether a character string (or vector) can be coerced to anumber (or numbers). The functions checkActiveDataSet, checkActiveModel,checkFactors, checkNumeric, checkTwoLevelFactors, and checkVariablescheck for the existence of objects and write an error message to the log if they are absent(or insufficiently numerous, in the case of different kinds of variables). The functioncheckReplace opens a dialog to query whether an existing object should be replaced.The function checkMethod, checks whether a method exists for a particular generic that isappropriate for a particular object. The function checkClass checks whether an object is ofa specific class. Both of these functions write error messages to the log if the condition fails.The function errorCondition reports an error to the user and (optionally) re-startsa dialog.

Information:

Several functions return vectors of object names: listAllModels,
listAOVModels, listCharacter, listDataSets, listDiscreteNumeric, listGeneralizedLinearModels,listFactors, listLinearModels, listMultinomialLogitModels,listNumeric, listProportionalOddsModels,listTwoLevelFactors,listVariables. The functions Predictors and Coefficients return information about the active model, or NULL if there is no active model. The functions activeDataSet and activeModel respectively report or set the activedata set and model.The function packageAvailable reports whether the named package isavailable to be loaded (or has possibly already been loaded).The function exists.method checks whether a method exists for a particular generic that isappropriate for a particular object, and returns TRUE or FALSE. The function is.SciViews always returns FALSEsince the SciViews GUI is no longer supported.

The function modelCapability() returns TRUE if there is an active statistical model and if it has the specified capability.For example, modelCapability("sum") returns TRUE if the model-capabilities table indicates that there's an applicablesummary() method for the active model. Otherwise, FALSE is returned. If the specified capability doesn't exist, a warning is printed.

Building dialog boxes:

Several functions simplify the process of constructing Tk dialogs:initializing a dialog box, initializeDialog, and completing the definition of a dialog box, dialogSuffix;a set of check boxes, checkBoxes;a set of radio buttons, radioButtons;a list box with associated scrollbars and state variable, variableListBox(and associated methods for the generic functionsgetFrame and getSelection);a drop-down "combo" box, variablecomboBox(with getFrame and getSelection methods);a button and subdialog for selecting a "grouping" variable, groupsBox;displaying the currently defined groups in a dialog, groupsLabel;a dialog-box structure for entering a model formula, modelFormula;a text box for entering a subsetting expression, subsetBox;OK, Cancel, and Help buttons for dialogs, OKCancelHelp, and subdialogs, subOKCancelHelp.The functions putDialog, getDialog, and varPosn support dialog-box memory—i.e., retainingselections across invocations of a dialog. The tkspinbox function is omitted from the tcltk package andmay be used to create a spinbox widget. The titleLabel function may be used to format a title label to usethe standard title label font and color.

“Themed” Tk widgets:

Tk 8.5 introduced so-called “themed” widgets, which look better than the traditional Tk widgets.Several functions, contributed by Brian Ripley, are written to access the new widgets by switching automaticallybetween the new and old widget sets depending upon the availability of the former: buttonRcmdr,to access either ttkbutton or tkbutton; labelRcmdr,to access either ttklabel or tklabel; ttkentry,to access either ttkentry or tkentry; ttkframe,to access either ttkframe or tkframe; ttkradiobutton,to access either ttkradiobutton or tkradiobutton; and ttkscrollbar,to access either ttkscrollbar or tkscrollbar.Note that the last four functions mask functions of the same names in the tcltk package.

‘Predicate’ functions:

A number of functions of the form nameP are ‘predicate’ functions, which returnTRUE or FALSE depending upon whether some condition obtains. For example, lmP()returns TRUE if there is an active model that is a linear model; and factorsP(2)returns TRUE if there are at least two factors in the active data set.WindowsP(), MacOSXP(), and X11P() return TRUE if the R Commander is running under Windows, Mac OS X, orX-Windows, consecutively.

Translating text:

The gettextRcmdr function simply passes its argument(s) togettext, adding the argument domain="R-Rcmdr" It is not meant to be used in plug-in packages.

Miscellaneous:

The function trim.blanks removes spaces from the beginning and end of a character string. The functioninstallPlugin installs an Rcmdr plug-in from a ZIP file or directory; this function may be used tocreate self-installing plug-ins in the form of packages. The function nobs returns the number ofobservations on which a statistical model is based. The function formulaFields returns informationabout the left-hand side, right-hand side, data, subset, and (for GLMs) family and link, of a model object.The function sortVarNames sorts variable names, including those containing numerals, into a more"natural" order than does the standard sort function. The function Library may be used toload packages; it checks whether a package is already loaded, and if not by default puts it in position4 on the search path. The function Coef, with several methods, returns the coefficients of a modelas a vector; the default method just calls coef.

The function RExcelSupported is used for the RExcel interface.

The function getCases takes a character string of case names or numbers as an argument and returns a character string that evaluates to an expression either to delete or retain these cases.

Some of these functions, marked # macro under Usage, are "macro-like" in their behaviour, in that theyexecute in the environment from which they are called. These were defined with an adaptation (used with permission) ofThomas Lumley's defmacro function, described in Lumley (2001), and are used in the R Commander to deal withscoping issues related to Tcl/Tk.

The tkfocus function is exported for historical reasons.

Author(s)

John Fox [emailprotected]

References

T. Lumley (2001) Programmer's niche: Macros in R.R News, 1(3), 11–13.

See Also

Command.

Description

This is a slightly modified version of the tkpager, changed to use the Rcmdr monospaced fontand a white background.

Usage

RcmdrPager(file, header, title, delete.file)

Arguments

See Also

tkpager

Description

The recode dialog is normally used to recode numeric variables and factorsinto factors, for example by combining values of numeric variables or levelsof factors. It may also be used to produce new numeric variables. The Rcmdrrecode dialog is based on the Recode function in thecar package.

Details

The name of each new variable must bea valid R object name (consisting only of upper and lower-case letters, numerals, andperiods, and not starting with a numeral).

Enter recode directives in the box near the bottom of the dialog. Directives are normally entered one per line, but may also be separated by semicolons. Each directive is of the form input = output (see the examples below). If an input value satisfies more than one specification, then the first (from top to bottom, and left to right) applies. If no specification is satisfied, then the input value is carried over to the result. NA is allowed on input and output. Factor levels are enclosed in double-quotes on both input and output.

Several recode specifications are supported:

a single value

For example, "missing" = NA.

several values separated by commas

For example, 7,8,9 = "high".

a range of values indicated by a colon

For example, 7:9 = "high". Thespecial values lo and hi may appear in a range. For example,lo:10=1. Note that these values are unquoted.

the special value else

everything that does not fit a previousspecification. For example, else=NA. Note that else matches all otherwise unspecified values on input, including NA.

If all of the output values are numeric, and the "Make new variable a factor" check box is unchecked, then a numeric result is returned.

If several variables are selected for recoding, then each is recoded using the same recodedirectives. In this case, the name entered in the box labelled"New variable name or prefix for multiple recodes" will be prefixed to the name of eachvariable being recoded. Setting an empty prefix (i.e., "") will cause the recoded variablesto replace the original variables.

As explained, = is used to separate old from new values, and : is used tospecify an interval (or range) of numeric values. It is possible to change these operatorsto other character strings, such as -> and ~ (tilde). This may be necessary, for example, if a factor to berecoded has =s or :s in its level (category) names.

Similarly, the dialog generates a call to the Recode function in the car package, which bydefault uses ; to separate recode specifications. The recode separator can also be changed, for example, to /.

Author(s)

John Fox [emailprotected]

See Also

Recode

Description

The one-way and two-way repeated-measures ANOVA/ANCOVA dialogs compute analysis of variance and analysis of covariance tables for one or two repeated-measures factors and a between-subjects linear model that can include both factors and covariates.

Details

The data are assumed to be in "wide" format, with repeated measures corresponding to distinct variables in the active data set. If the data are in "long" format, they can be reshaped to wide format via the Data > Active data set > Reshape data set from long to wide format dialog.

The model is specified in the Design tab in two parts:

1. The within-subjects design is defined by using the drop-down lists to select the variables that correspond to the levels of the within-subjects factor (in the case of one repeated-measures factor) or the combinations of levels of the within-subjects factors (in the case of two repeated-measures factors, organized as a two-way table). The user can also name the levels of the within-subjects factor(s) and the factor or factors themselves.

2. The between-subjects model is specified exactly as for a linear model (see linearModel). The model is then fit as a multivariate linear model with the repeated measures as response variables.

In the Options tab, the user can select either multivariate tests (using one of four test statistics) or univariate tests. The tests are performed by the Anova function in the car package.

The means of the repeated measures can optionally be plotted as a function of within- and between-subjects factors, and the means and standard deviations can be printed, using the repeatedMeasuresPlot function in the RcmdrMisc package.

Author(s)

John Fox [emailprotected]

References

John Fox, Michael Friendly, and Sanford Weisberg,"Hypothesis Tests for Multivariate Linear Models Using the car Package",The R Journal, 5 (1) 39–52, 2013.

See Also

linearModel, Anova, repeatedMeasuresPlot

Description

There are two dialogs for "reshaping" the active data set: (1) from "long" to "wide" format (using the reshapeL2W in the RcmdrMisc package); and (2) from "wide" to "long" format (using the reshapeW2L in the RcmdrMisc package).

Details

These dialogs are designed for handling regularly structured repeated-measures data, where each subject (independent unit of observation) is measured on several occasions or under several different conditions. The occasions may have either a one-dimensional structure (corresponding to the levels of one repeated-measures or within-subjects factor) or a two-dimensional structure (corresponding to two crossed repeated-measures factors).

Data in "wide" format have one row for each subject, with the repeated measures in different columns (variables). Data in "long" format have several rows for each subject, with one column (variable) for the response; the levels of the repeated-measures factor (or combinations of levels for two repeated-measures factors) correspond to distinct rows. The within-subjects factor or factors appear as different columns in the long data, as do the between-subjects factors; the former vary within subjects, while the latter are invariant within subjects; and an ID variable identifies the rows of the data set belonging to each subject.

Data in wide format are suitable for analysis by the R Commander repeated-measures ANOVA/ANCOVA dialogs (see RepeatedMeasuresDialogs), while data in long format are suitable for analysis by the R Commander mixed-models dialogs.

The Reshape Data Set from Long to Wide Format dialog is largely self-explanatory: The user selects the variable that identifies subjects (i.e., the ID variable); one or two within-subjects factors; the variable or variables that vary by occasion (there is typically just one, the response variable); and any variables in the long data set that are to be excluded from the wide data set.

The Reshape Data Set from Wide to Long Format dialog is more complex. There are three tabs: A tab to specify one repeated-measures factor; a tab to specify two crossed repeated-measures factors; and an options tab. The user specifies either one or two repeated-measures factors, not both. The within-subjects factor or factors are defined by using drop-down lists to select the variables that correspond to the levels of the within-subjects factor (in the case of one repeated-measures factor) or the combinations of levels of the within-subjects factors (in the case of two repeated-measures factors, organized as a two-way table). The user can also name the levels of the within-subjects factor(s) and the factor or factors themselves.

Author(s)

John Fox [emailprotected]

See Also

reshapeL2W, reshapeW2L, RepeatedMeasuresDialogs.

Description

This dialog creates a ‘.Rprofile’ file, by default in the current directory, adding to it the currentR Commander options, set, e.g., in the Options dialog. If R is subsequently restarted in this directory, and the Rcmdr package loaded,then the current R Commander options will be applied. The current directory will typically, thoughnot necessarily, be your home directory — for example, for Windows users, your Documentsdirectory.

If a ‘.Rprofile’ file already existsin the current directory, then the R Commander options are added to it at the end of the file,after removing R Commander options previously generated by an earlier invocation of this dialog.

The contents of the ‘.Rprofile’ file may be edited before you save the file.If you want to start the R Commander automatically when R starts, uncomment (remove the #s from)the four lines

 # local({ # old <- getOption('defaultPackages') # options(defaultPackages = c(old, 'Rcmdr')) # })

See Startup for a description of the ‘.Rprofile’file and the R startup process in general.

See Commander for a description of the various R Commander options.

Usage

saveOptions() # typically through the R Commander Tools menu

Author(s)

John Fox [emailprotected]

See Also

Commander, Startup

Description

This dialog sets up a call to the scatter3d function to draw athree-dimensional scatterplot, and optionally to Identify3d to labelpoints interactively with the mouse.

Details

The explanatory variables provide the "horizontal" and "out-of-screen" axes of the scatterplot,the response variable provides the "vertical" axis.

Data points are represented as spheres or points,depending upon the number of observations.

Several regression surfaces can beplotted: a linear least-squares surface; a full quadratic least-squares surfacewith squared and cross-product terms; a "smooth" regression surface — either asmoothing spline, if no degrees of freedom are specified (in which case thegam function selects the df by generalized cross validation),or a fixed-df regression spline; an additive-regression surface (also fit by gam),with either smoothing spline or regression spline components (again selected accordingto the specification of degrees of freedom). If only one surface is fit, then residuals areplotted as red (negative) and green (positive) lines from the surface to the points.

You can specify a factor defining groups by pressing the Plot by groups button. Aseparate surface or set of surfaces is plotted for each level of the groups factor. Thesesurfaces can be constrained to be parallel.

The completed plot can be manipulated with the mouse: Click, hold, drag the left mouse button to rotatethe display; click, hold, and drag the right button (or centre button on a three-button mouse)to zoom in and out.

If the box labelled Identify observations with mouse is checked, you may usethe mouse to identify points interactively: Press the right mouse button (or the centre buttonon a three-button mouse), drag a rectangle around the points to be identified, and release the button. Repeat this procedure for each point orset of "nearby" points to be identified. To exit from point-identification mode,right-click (or centre-click) in an empty region of the plot.

Points may also be identified subsequently by selecting Identify observations with mousefrom the R Commander 3D graph menu: As above, click and drag the left mouse button torotate the display, and click and drag the right (or centre) button to identify points.

Author(s)

John Fox [emailprotected]

See Also

scatter3d, Identify3d, rgl-package, gam

Description

The R Commander script editor is meant to edit scripts in textwidgets, such as the R Commander R Markdown and knitrdocument tabs.

Details

Saving the document, either via the File menu or pressing the OK buttoncloses the editor and modifies the content of the correspondingR Markdown or knitr tab. Closing the editor without saving, byselecting Cancel from the file menu, pressing the Cancel button ordestroying the window discards changes to the document. You may also save your editsto the R Markdown or knitr tab without closing the editor. Compiling the documentinto a report also saves the current edits.

The editor is a “non-modal” dialog, and so may remain open when you work.

Author(s)

John Fox [emailprotected]

See Also

using R Markdown; using knitr.