latticist: A Lattice GUI
In 11010tianyi/latticist: A graphical user interface for exploratory visualisation

Description Usage Arguments Details Value Author(s) References See Also Examples

Interactively explore a data set using Lattice displays.

latticist(dat,
          spec = list(),
          reorder.levels = !is.table(dat),
          ...,
          use.playwith = latticist.getOption("use.playwith"))

`dat`	a `data frame` (with numeric and/or categorical variables), or a `table`.
`spec`	a list specifying the initial latticist display. See `latticistCompose`.
`reorder.levels`	if `TRUE`, change the ordering of levels of factor variables so they are in order of frequency. Note that this does not change the data, only its internal representation: see `reorder`. This is recommended for effective graphic displays. Factor variables of class `ordered` are not reordered. Therefore you should ensure that any variables whose levels have an inherent order are of class `ordered`: see `as.ordered`. Note: the result of `cut` is by default NOT `ordered`! In addition, numeric variables with discrete values in (0, 1, -1) are converted to factors.
`use.playwith`	whether to launch as a toolbar for `playwith` (requires the playwith package), or a generic gWidgets interface.
`...`	further arguments specific to the interface type. For the `playwith` interface, these are passed to `playwith`. For the `gWidgets` interface, these are: `width = 450, height = 450` size of the graphics device in pixels. Note, these are ignored if not using an embedded device. `pointsize = 12` text size on the graphics device.

Latticist is a graphical user interface for exploratory visualisation. It is primarily an interface to the Lattice graphics system (from the lattice package), but also produces displays from the vcd package for categorical data.

Given a multivariate dataset (either a data frame or a table), Latticist attempts to produce useful displays based on the properties of the data. The user chooses variables or expressions for the plot axes, for grouping, conditioning and subsetting. Some hypervariate displays are also available.

A minimal graphical user interface is available, built on the gWidgets package. This requires one of the "toolkit implementations" to installed: gWidgetstcltk, gWidgetsRGtk2 or gWidgetsrJava. Note that gWidgetsrJava is currently broken (as of version 0.0-13).

The fastest way to start, without any external system requirements, is to install.packages("gWidgetstcltk"). However, that does not support an embedded graphics device, so the plots will appear in a separate window.

Alternatively, Latticist can be run as a toolbar extension to playwith. This brings many extra features, such as dynamic zooming, identifying data points, linked brushing, etc. Note that the playwith package requires RGtk2 and, therefore, the GTK+ libraries.

To enable all types of graphic displays, install the hexbin and deldir packages.

The playwith method invisibly returns the playState object representing the plot window. One can close it with playDevOff() or dev.off().

The gWidgets method invisibly returns the gwindow object. One can close it with dispose().

Felix Andrews felix@nfrac.org

For an excellent introduction to and coverage of Lattice:

Sarkar, Deepayan (2008) “Lattice: Multivariate Data Visualization with R”, Springer. ISBN: 978-0-387-75968-5 http://lmdvr.r-forge.r-project.org/

The mosaic displays and extensions from vcd are well described in:

David Meyer, Achim Zeileis, and Kurt Hornik (2006). “The Strucplot Framework: Visualizing Multi-Way Contingency Tables with vcd”. Journal of Statistical Software, 17(3), 1-48. http://www.jstatsoft.org/v17/i03/

latticistCompose, latticist.options, Lattice, lattice.demo in the TeachingDemos package.

if (interactive()) {
options(device.ask.default = FALSE)

## Not run: 
## data frame example:
latticist(iris)

## table example:
latticist(Titanic, spec = list(groups = "Survived"))

## End(Not run)

## The GUI comes in three flavours:

if (require("gWidgetstcltk") &&
    isTRUE(gconfirm("Show tcl/tk-based GUI?")))
{
   options(guiToolkit = "tcltk")
   latticist(CO2, use.playwith = FALSE)
}

if (require("gWidgetsRGtk2") &&
    isTRUE(gconfirm("Show RGtk2-based GUI?")))
{
   options(guiToolkit = "RGtk2")
   latticist(CO2, use.playwith = FALSE)
}

if (require("playwith") &&
    isTRUE(gconfirm("Show playwith-based GUI?")))
{
   latticist(CO2, use.playwith = TRUE)
}

}