Learn R Programming

loa (version 0.3.1.1)

1.1.loaPlot: loaPlot, XYZ plots for lattice

Description

loaPlot is a standard XYZ plotting function, where X and Y are the axes of a conventional XY plot and Z is an element (or elements if supplied in the form Z1 + Z2 + Z3...) visualized at associated XY coordinates. By default, loaPlot links Z to plot point size and color to generate a bubbleplot style output, or using modified plot calls other plot types.

Usage

loaPlot(x, data = NULL, panel = panel.loaPlot, 
        ..., local.scales = FALSE, reset.xylims = TRUE, 
        load.lists = NULL, by.group = NULL, by.zcase = NULL, 
        preprocess = TRUE)

#standard panels

panel.loaPlot(..., loa.settings = FALSE) panel.loaPlot2(..., loa.settings = FALSE) panel.loa(..., loa.settings = FALSE)

#grids

panel.loaGrid(grid.x = NULL, grid.y = NULL, xlim = NULL, ylim = NULL, ..., grid = NULL, panel.scales = NULL)

Value

loaPlot returns a trellis object, much like a conventional lattice

plot function.

Arguments

x

A formula with the general structure z ~ x * y | cond applied like in the lattice function levelplot or a matrix. For a formula, x and y are the horizontal and vertical axes, z is any additional information to be used in point, symbol, surface or glyph generation, and cond is any additional conditioning to be applied. x and y are required elements; z and cond are typically optional.
(Note: this element of the plot is handled by formulaHandler).

data

If supplied, the assumed source of elements of x, typically a data.frame.

panel

panel is the function to be used when generating the content of the individual panels within the lattice plot. By default, this is the loa panel function panel.loaPlot.

...

Additional arguments are passed on to related functions.

For loaPlot these are colHandler, cexHandler and the function set by panel.

This mechanism provides access to most common plot parameters, e.g. col, pch, and cex for plot symbol color, type and size, respectively. By default, both data point color and size are z-scaled for loaPlot. If z is supplied, and cex and col are not set by the user in the plot command, these plot properties are managed by cexHandler and colHandler, respectively. cexHandler and colHandler arguments can be also be passed directly as part of the loaPlot command to fine-tune these, e.g. cex.range to change the cex range that z values are scaled to and col.region to change the color range that is applied to z when coloring points. See associated Help documents for further information.

local.scales

For loaPlot only, logical. If TRUE, this removes the standard lattice axis from the plot. It is intended to be used with panel functions which generate their own axes or have no axes.

reset.xylims

For loaPlot only, logical or character vector. If a logical, if the panel outputs are preprocessed (using panelPal), should the x and y limits be reset? If a character vector, one or more terms controlling post-processing plot range management: refit.xylims, equivalent to reset.xylims = TRUE; and max.xylims, to reset both x and y ranges to maximum.
(Note: If xlim or ylim are supplied in the plot call, these will typically override all reset.xylims settings.)

load.lists

For loaPlot only, character vector. In-development alternative to list based arguments. This option identifies plot call arguments that loaPlot should manage using listLoad. See associated help documentation for further details.

by.group, by.zcase

For loaPlot only. Arguments for routine by group and by zcase handling of plot inputs.
Important: These are current under review.

preprocess

For loaPlot only, logical, passed to panelPal. If TRUE, and used with a correctly configured panel function, this processes the plot input before generating the plot. This means color scales in the different plot panels and the key are automatically aligned and the associated trellis object output contains the panel function outputs rather than the inputs. See panelPal Help documents for further information.

loa.settings

For panel... functions only, logical, passed to panelPal to manage plot reworking. See associated Help documents for further information.

grid.x, grid.y, xlim, ylim, grid, panel.scales

For panel.loaGrid only, grid settings, typically recovered by loaPlot.

Author

Karl Ropkins

Details

loaPlot provides lattice-style conditioning/handling for a range of commonly used XYZ plotting options. It is perhaps easiest pictured as a 'mid point' alternative somewhere between the standard lattice plot functions xyplot and levelplot.

The default form of the plot uses an extension of the subscripting methods described by Deepayan Sarkar in Chapter 5 of Lattice (see sections on scatterplots and extensions). The default output is a bubble plot (see example 1 below).

References

These functions make extensive use of code developed by others.

lattice: Sarkar, Deepayan (2008) Lattice: Multivariate Data Visualization with R. Springer, New York. ISBN 978-0-387-75968-5

RColorBrewer: Erich Neuwirth <erich.neuwirth@univie.ac.at> (2011). RColorBrewer: ColorBrewer palettes. R package version 1.0-5. http://CRAN.R-project.org/package=RColorBrewer

See Also

In loa: panelPal

In other packages, see

lattice: xyplot; and levelplot.

Examples

Run this code

## Example 1
## Basic usage

loaPlot(Ozone~Solar.R*Temp|Wind>8, 
        data=airquality, 
        col.regions="Blues")

# Notes:
# Formula structure z ~ x * y |cond like levelplot.
# Data (and groups) assignment like in standard lattice plots.
# By default z is linked to col and cex. 
# Unless overridden by user inputs or group or zcase setting. 
# Plot passed via ...Handler functions to provide shortcut plot 
# reworking, e.g. here colHandler handles color scales 
# using col.region to generate a color range. 
# (Here, arguments like "Blues" and c("green", "red") are 
# allowed and handled using functions in the RColorBrewer 
# package.)

# Formula stuctures:
# ~ x * y                like xyplot y ~ x
# ~ x * y | cond         like xyplot y ~ x | cond
# z ~ x * y              like xyplot y ~ x, col=f(z), cex=f(z)
# z ~ x * y | cond       like xyplot y ~ x | cond, col=f(z), cex=f(z)
# z ~ x * y, groups = g  like xyplot y ~ x, groups=g, cex=f(z)
# z1 + z2 ~ x * y        (zcases)
# etc


## Example 2 
## Basic modifications

loaPlot(Ozone~Solar.R*Temp, groups=airquality$Wind>8, 
        data=airquality) 

# When groups are applied, by default group id is linked to col.
# The follow example illustrates three options:

loaPlot(Ozone~Solar.R*Temp, groups=airquality$Wind>8, 
        data=airquality,  
        group.args=c("pch"), pch=c(1,4), 
        col="blue")

# notes:
# Here, group.args is used to change the default group arguments.
# (So, pch rather than col is used to identify groups.)
# pch is then assigned by group rather than by (x,y) case or z case. 
# (See panelPal Help further further details of assignments in loa.)
# col supplied by the user supercedes the default z linkage.
# (So, here cex remains z scales but col is fixed as blue.)


## Example 3 
## Key handling

loaPlot(Ozone~Solar.R*Temp, data=airquality, 
        col.regions=c("green", "red"))

# Key settings are by the key argument (as in lattice) 
# or key... arguments via keyHandler and listLoad, so e.g.:

loaPlot(Ozone~Solar.R*Temp, data=airquality, 
        col.regions=c("green", "red"), 
        key.fun = draw.loaColorKey)

# Notes:
# By default the loaPlot uses draw.loaPlotZKey to generate 
# its color key unless an alternative is supplied via key.fun.
# (Here, the draw.colorKey wrapper draw.loaColorKey is used to 
# generate a color bar similar to that in levelplot.)



## Example 4
## panels

loaPlot(Ozone~Solar.R*Temp|Wind>8, data=airquality, 
        col.regions="Reds")

# The combined use of loaPlot, panelPal and appropriately configured 
# panel functions provides automatical handling of a range of plot 
# elements, e.g.:

loaPlot(Ozone~Solar.R*Temp|Wind>8, data=airquality, 
        col.regions="Reds", panel=panel.binPlot)

# Notes:
# Here, the choice of default key is set by the panel... function;
# the panel by default bins data by location and for each bin cell
# calculates the mean Ozone concentration just like a standard 
# lattice panel would, but it also tracks these values (calculated 
# within the panels) and scales panel-to-panel and panel-to-key 
# so users do not have to do that retrospectively; and, finally, 
# it retains in-panel calculations so users can recover them.
# (See associated helps for further details: ?panelPal about methods; 
# and ?panel.binPlot about the panel function.)
 

Run the code above in your browser using DataLab