R/plotting_functions.R
In rettopnivek/rtplots: Plotting functions for simple choice and response time data.
#----------------------------------#
# Functions for plotting data from #
# simple choice and response time #
# tasks #
#----------------------------------#
# Useful functions for package creation
# library(devtools)
# library(roxygen2)
# * Incomplete documentation
# Index
# Lookup - 01: rtplots*
# Lookup - 02: is.rtplots
# Lookup - 03: lines.rtplots
# Lookup - 04: points.rtplots
# Lookup - 05: plot.rtplots
# Lookup - 06: uncertainty*
# Lookup - 07: draw_ui
# Lookup - 08: add_to_rtplot*
# Lookup - 09
# Lookup - 01
#' Creates a 'rtplots' Object
#'
#' Creates a 'rtplots' object from a data frame of
#' response time and choice/accuracy data, allowing
#' for subsequent plotting.
#'
#' @param df a data frame that includes (at a minimum) a column of
#' response times.
#' @param label a character string giving the column name for
#' the response time variable, and optionally the choice/accuracy
#' variable, a grouping factor variable (e.g., subjects), and
#' finally an additional covariate (e.g., conditions). If
#' \code{NULL}, the algorithm will attempt to locate response
#' time and choice/accuracy variables via a set of default names.
#' @param keep a logical vector, indicating which rows should be
#' kept when extracting the variables.
#' @param type the type of function to compute. Options are
#' \itemize{
#' \item \code{CDF}: the cumulative distribution function.
#' \item \code{QPE}: the quantile function.
#' \item \code{PDF}: the probability density function.
#' \item \code{PVT}: probability versus time.
#' }
#' @param level an optional index indicating the specific level
#' of the grouping factor to consider. This allows, for instance,
#' selecting a single subject.
#' @param ... additional parameters for the density, distribution,
#' and quantile functions.
#'
#' @details
#'
#' Forthcoming
#'
#' @return An object of class 'rtplots', a list consisting of...
#' \itemize{
#' \item \code{pd}; a data frame with the x and y values
#' to plot and their associated choice/accuracy value.
#' \item \code{xm}; If relevant, a matrix whose columns
#' contain the individual x-axis values of the function
#' which were collapsed over the grouping factor.
#' \item \code{ym}; If relevant, a matrix whose columns
#' contain the individual y-axis values of the function
#' which were collapsed over the grouping factor.
#' \item \code{npfd}; a data frame with the number of
#' observations, proportions, and frequencies based on
#' the choice/accuracy values.
#' \item \code{type}; the type of function that was applied.
#' \item \code{x}; the range of x-axis values.
#' \item \code{y}; the range of y-axis values.
#' }
#'
#' @references
#'
#' Forthcoming
#'
#' @examples
#' # Load in example dataset
#' data("priming_data")
#' lbl = c( 'RT', 'Accuracy', 'Subjects' )
#' keep = priming_data$Condition == 0
#' obj = rtplots( priming_data, label = lbl, keep = keep )
#'
#' @export
rtplots = function( df, label = NULL, keep = NULL,
type = 'CDF', level = NULL, ... ) {
# Check input
if ( !is.data.frame( df ) )
stop( 'Input variable must be a data frame' )
output = extract_var( df, label, keep, type )
# Sort RTs into ascending order
output = sort_input( output, level )
# Compute proportions and frequencies
npfd = extract_npfd( output )
if ( check_for_cdf_type(type) ) {
### CDF
# Generate output
tmp = create_cdf_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'CDF'
} else if ( check_for_qpe_type(type) ) {
### QPE
# Generate output
tmp = create_qpe_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'QPE'
} else if ( check_for_pdf_type(type) ) {
### PDF
# Generate output
tmp = create_pdf_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'PDF'
} else if ( check_for_pvt_type(type) ) {
### PVT
# Generate output
tmp = create_pvt_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'PVT'
} else {
err = 'That function type is unknown.'
stop( err, call. = FALSE )
}
# Plotting range
output$x = range( na.omit( output$pd$x ) )
output$y = range( output$pd$y )
class( output ) = 'rtplots'
return( output )
}
# Lookup - 02
#' @rdname rtplots
#' @export
is.rtplots = function(x) inherits(x, "rtplots")
# Lookup - 03
#' Lines Method for 'rtplots' Object
#'
#' Provides a method to 'rtplots' objects that allows
#' for drawing line segments on an existing plot.
#'
#' @param object a 'rtplots' object.
#' @param val the choice or accuracy value whose observations
#' are to be plotted.
#' @param ind a index to select observations for a specific
#' choice/accuracy value.
#' @param flip logical; if true, flips the y-axis values to
#' be negative.
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{lines}}.
#'
#' @export
lines.rtplots = function( object, val = NULL,
ind = 1, flip = F, ... ) {
# If it isn't a probability versus time plot
if ( object$type != 'PVT' ) {
if ( !is.null( val ) ) {
vl = val == object$pd$v
} else {
vl = object$val[ ind ] == object$pd$v
}
if ( !flip ) {
lines( object$pd$x[vl], object$pd$y[vl], ... )
} else {
lines( object$pd$x[vl], -object$pd$y[vl], ... )
}
} else { # Special options for PvT figures
# Extract choice value
if ( is.null( val ) ) val = object$val[ ind ]
# Check for multiple RTs per probability
dmn = dim( object$pd$y )
# If the points aren't flipped about the x-axis
if ( !flip ) {
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:dmn[2] ) {
x = object$pd$x[vl]
y = object$pd$y[vl,i]
ord = order(x)
lines( x[ord], y[ord], ... )
}
} else { # Flip
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:dmn[2] ) {
x = object$pd$x[vl]
y = object$pd$y[vl,i]
ord = order(x)
lines( x[ord], -y[ord], ... )
}
}
}
}
# Lookup - 04
#' Points Method for 'rtplots' Object
#'
#' Provides a method to 'rtplots' objects that allows
#' for drawing points on an existing plot.
#'
#' @param object a 'rtplots' object.
#' @param val the choice or accuracy value whose observations
#' are to be plotted.
#' @param ind a index to select observations for a specific
#' choice/accuracy value.
#' @param flip logical; if true, flips the y-axis values to
#' be negative.
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{points}}.
#'
#' @export
points.rtplots = function( object, val = NULL,
ind = 1, flip = F, ... ) {
# If it isn't a probability versus time plot
if ( object$type != 'PVT' ) {
# If a choice value is given
if ( !is.null( val ) ) {
vl = val == object$pd$v
} else {
vl = object$val[ ind ] == object$pd$v
}
# If the points aren't flipped about the x-axis
if ( !flip ) {
points( object$pd$x[vl], object$pd$y[vl], ... )
} else { # Flip
points( object$pd$x[vl], -object$pd$y[vl], ... )
}
} else { # Special options for PvT figures
# Extract choice value
if ( is.null( val ) ) val = object$val[ ind ]
# Check for multiple RTs per probability
dmn = dim( object$pd$y )
# If the points aren't flipped about the x-axis
if ( !flip ) {
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:length(vl) ) {
x = rep( object$pd$x[vl[i]], dmn[2] )
y = object$pd$y[vl[i],]
points( x, y, ... )
}
} else { # Flip
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:length(vl) ) {
x = rep( object$pd$x[vl[i]], dmn[2] )
y = object$pd$y[vl[i],]
points( x, -y, ... )
}
}
}
}
# Lookup - 05
#' Plot Method for 'rtplots' Object
#'
#' Creates a blank plot based on a 'rtplots' object.
#'
#' @param x a 'rtplots' object.
#' @param y the y-axis boundaries.
#' @param inc the interval for the ticks on the
#' x and y-axes (depending on the plot).
#' @param type if equal to \code{'blank'}, a completely
#' blank plot is generated. Otherwise, converted to \code{'n'}.
#' @param xlab a character string for the label on the x-axis.
#' @param ylab a character string for the label on the y-axis.
#' @param bty the type of box to draw around the plot. See
#' \code{\link[graphics]{par}}.
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{par}}.
#'
#' @export
plot.rtplots = function( x,
y = NULL,
inc = NULL,
type = NULL,
xlab = NULL,
ylab = NULL,
bty = NULL, ... ) {
# Extract 'rtplots' object
object = x
if ( object$type == 'CDF' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
xl = lower_upper( inc[1], object$pd$x )
if ( is.null( y ) ) yl = c(0,1) else yl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Time' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Distribution function'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
if ( object$type == 'PDF' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
xl = lower_upper( inc[1], object$pd$x )
if ( is.null( y ) ) yl = c(0,2) else yl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Time' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Density function'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
if ( object$type == 'QPE' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
xl = lower_upper( inc, na.omit( object$pd$x ) )
if ( is.null( y ) ) yl = c(0,1) else yl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Time' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Quantile function'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
if ( object$type == 'PVT' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
yl = lower_upper( inc, as.vector( na.omit( object$pd$y ) ) )
if ( is.null( y ) ) xl = c(0,1) else xl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Probability' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Time'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
}
# Lookup - 06
#' Computes Uncertainty Intervals for 'rtplots' Objects
#'
#' A function that computes a specified uncertainty
#' interval (such as \eqn{\pm} 2 standard errors from
#' the mean) for a 'rtplots' object.
#'
#' @param object a 'rtplots' object.
#' @param f a function to compute the boundaries of
#' the uncertainty interval given the desired
#' range of coverage (takes at least two parameters,
#' a vector \code{x} and a range \code{interval}).
#' @param alpha the desired width of coverate for the
#' given uncertainty interval.
#' @param ... additional parameters for the
#' function \code{f}.
#'
#' @return Forthcoming
#'
#' @export
uncertainty = function( object, f = NULL,
alpha = .95, ... ) {
# If no function is provided, assume uncertainty
# intervals should be calculated using the
# standard error of the mean
if ( is.null( f ) ) {
f = function(x,interval) {
s = sd( x )/sqrt( length(x) ) # Standard error
m = mean(x)
return( c( m - abs( qnorm(interval[1]) )*s,
m + abs( qnorm(interval[2] ) )*s ) )
}
}
# Determine lower and upper boundaries
interval = numeric(2);
interval[1] = (1 - alpha)/2
interval[2] = interval[1] + alpha
# Initialize output
ud = NULL
pd = NULL
# If there is more than one group
if ( object$n_g > 1 ) {
l = nrow( object$pd )
ud = data.frame( x = rep( NA, l * 3 ),
y = rep( NA, l * 3 ),
loc = rep( NA, l * 3 ),
v = rep( NA, l * 3 ),
i = rep( NA, l * 3 ) )
# Loop over choice/accuracy values
sta = 1; end = 0
for ( i in 1:(object$n_v) ) {
# Compute uncertainty intervals
if ( !is.null( object$xm ) ) {
xui = apply( object$xm[[i]], 2, f, interval = interval,
... )
} else xui = NULL
if ( !is.null( object$ym ) ) {
yui = apply( object$ym[[i]], 2, f, interval = interval,
... )
} else yui = NULL
sel = object$val[ i ] == object$pd$v
# Center values
end = end + sum( sel )
ud$x[ sta:end ] = object$pd$x[ sel ]
ud$y[ sta:end ] = object$pd$y[ sel ]
ud$v[ sta:end ] = object$pd$v[ sel ]
ud$loc[ sta:end ] = 'center'
ud$i[ sta:end ] = 1:sum(sel)
sta = end + 1
# Lower boundary
end = end + sum( sel )
if ( is.null( xui ) )
ud$x[ sta:end ] = object$pd$x[ sel ] else
ud$x[ sta:end ] = xui[1,]
if ( is.null( yui ) )
ud$y[ sta:end ] = object$pd$y[ sel ] else
ud$y[ sta:end ] = yui[1,]
ud$v[ sta:end ] = object$pd$v[ sel ]
ud$loc[ sta:end ] = 'lower'
sta = end + 1
# Upper boundary
end = end + sum( sel )
if ( is.null( xui ) )
ud$x[ sta:end ] = object$pd$x[ sel ] else
ud$x[ sta:end ] = xui[2,]
if ( is.null( yui ) )
ud$y[ sta:end ] = object$pd$y[ sel ] else
ud$y[ sta:end ] = yui[2,]
ud$v[ sta:end ] = object$pd$v[ sel ]
ud$loc[ sta:end ] = 'upper'
sta = end + 1
}
pd = data.frame( xl = rep( NA, l * 2 ),
xu = rep( NA, l * 2 ),
yl = rep( NA, l * 2 ),
yu = rep( NA, l * 2 ) )
pd$xl = c( ud$x[ ud$loc == 'lower' ],
ud$x[ ud$loc == 'center' ] )
pd$xu = c( ud$x[ ud$loc == 'upper' ],
ud$x[ ud$loc == 'center' ] )
pd$yl = c( ud$y[ ud$loc == 'center' ],
ud$y[ ud$loc == 'lower' ] )
pd$yu = c( ud$y[ ud$loc == 'center' ],
ud$y[ ud$loc == 'upper' ] )
pd$v = rep( ud$v[ ud$loc == 'center' ], 2 )
pd$i = rep( ud$i[ ud$loc == 'center' ], 2 )
}
out = list(
pd = pd,
ud = ud,
xm = object$xm,
ym = object$ym,
val = object$val
)
# Define new class
class( out ) = 'rtplots_ui';
return( out )
}
# Lookup - 07
#' Draw Uncertainty Intervals
#'
#' A convenience function that draws uncertainty
#' intervals computed from a 'rtplots' object onto
#' an already existing plot.
#'
#' @param ui a 'rtplots_ui' object.
#' @param val the choice/accuracy value for which to
#' draw the uncertainty intervals.
#' @param ind the index of the desired choice/accuracy
#' value over which to draw uncertainty intervals.
#' @param flip logical; if \code{true} the y-axis
#' values are flipped to be negative.
#' @param type the type of method used to draw the
#' uncertainty interval. Options are
#' \itemize{
#' \item \code{segments}.
#' \item \code{arrows}.
#' \item \code{polygon}.
#' }
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{segments}},
#' \code{\link[graphics]{arrows}}, and
#' \code{\link[graphics]{polygon}}.
#'
#' @export
draw_ui = function( ui, val = NULL, ind = 1,
flip = F, type = 'segments', ... ) {
if ( !is.null( val ) ) {
vl = val == ui$pd$v
} else {
vl = ui$val[ ind ] == ui$pd$v
}
if (!flip) w = 1 else w = -1
segment_names = c( 'segments', 's', 'seg' )
if ( type %in% segment_names ) {
segments( ui$pd$xl[vl], w*ui$pd$yl[vl],
ui$pd$xu[vl], w*ui$pd$yu[vl], ... )
}
polygon_names = c( 'polygon', 'p', 'poly' )
if ( type %in% polygon_names ) {
polygon( c( ui$pd$xl[vl], rev( ui$pd$xu[vl] ) ),
c( ui$pd$yu[vl], rev( ui$pd$yl[vl] ) ),
... )
}
arrow_names = c( 'arrows', 'a', 'arr' )
if ( type %in% arrow_names ) {
arrows( ui$pd$xl[vl], w*ui$pd$yl[vl],
ui$pd$xu[vl], w*ui$pd$yu[vl], ... )
}
}
# Lookup - 08
#' Add Points or Lines to a RT Plot
#'
#' A convenience function that draws points or
#' line segments onto an already existing plot. Useful
#' for adding summary statistics (e.g., the mean or
#' median).
#'
#' @param object a rtplots object or a list with the
#' x and y-axis values to plot.
#' @param T_x the test statistic(s) to be added to the
#' existing figure.
#' @param T_g the test statistic to use when collapsing
#' over a grouping factor.
#' @param val Forthcoming.
#' @param ind Forthcoming.
#' @param type Forthcoming.
#' @param out Forthcoming.
#' @param ... Forthcoming.
#'
#' @export
add_to_rtplot = function( object, T_x = mean, T_g = mean, val = NULL,
ind = 1, type = 'p', out = T, ... ) {
# Function to calculate values for y-axis
# using linear interpolation
f = function( x ) {
ind = max( which( xa < x ) )
if ( ind == -Inf ) {
out = ya[1];
} else {
if ( ind == length(xa) ) {
out = ya[ length(xa) ]
} else {
pts = c( ind, ind + 1 )
out = ln_interp( x, ya[ pts ], xa[ pts ] )
}
}
return( out )
}
if ( is.rtplots( object ) ) {
# If it isn't a probability versus time plot
if ( object$type != 'PVT' ) {
# If there is no grouping factor
if ( object$n_g <= 1 ) {
if ( !is.null( val ) ) {
vl = val == object$ad$v
vly = val == object$pd$v
} else {
vl = object$val[ ind ] == object$ad$v
vly = object$val[ ind ] == object$pd$v
}
ts = T_x( object$ad$t[vl] )
# Extract plotting values
xa = object$pd$x[vly] # x-axis
ya = object$pd$y[vly] # y-axis
# Determine y-axis values
ny = sapply( ts, f )
output = list( x = ts, y = ny )
} else {
if ( !is.null( val ) ) {
vl = val == object$ad$v
vly = val == object$pd$v
} else {
vl = object$val[ ind ] == object$ad$v
vly = object$val[ ind ] == object$pd$v
}
dmn = length( T_x( object$ad$t[vl] ) )
ts_all = by( object$ad$t[vl], list( object$ad$g[vl] ), T_x )
ts_all = matrix( unlist( ts_all ), length( ts_all ), dmn, byrow = T )
ts = apply( ts_all, 2, T_g )
# Extract plotting values
xa = object$pd$x[vly] # x-axis
ya = object$pd$y[vly] # y-axis
# Determine y-axis values
ny = sapply( ts, f )
output = list( x = ts, y = ny )
}
}
} else {
pass = F
if ( is.list( object ) ) {
if ( all( c('x','y') %in% names( object ) ) ) {
pass = T
output = object
}
if (!pass) stop( 'Please pass in an "rtplots" object',
.call = FALSE )
}
}
if ( type == 's' | type == 'segments' ) {
segments( output$x, 0, output$x, output$y, ... )
}
if ( type == 'p' | type == 'points' ) {
points( output$x, output$y, ... )
}
if ( out ) return( output )
}
# Lookup - 09
#' Create a Probability versus Time Figure
#'
#' A convenience function that draws points, lines, or
#' line segments for a probability versus time figure
#' onto an already existing plot. Allows fine-tune
#' control of plotting parameters such as point type,
#' size, etc.
#'
#' @param object a rtplots object.
#'
#'
#' @export
pvt_plot = function( object, keep = NULL, ind = NULL, val = NULL, type = 'p',
pch = NULL, cex = NULL, col = NULL,
bg = NULL, lwd = NULL ) {
if ( is.rtplots( object ) ) {
if ( object$type == 'PVT' ) {
# If a choice value is given
if ( !is.null( val ) ) {
vl = val == object$pd$v
} else if ( !is.null( ind ) ) {
vl = object$val[ ind ] == object$pd$v
} else {
vl = rep( T, nrow( object$pd ) )
}
# Extract relevant data
if ( is.null( keep ) ) keep = rep( T, nrow( object$pd ) )
# Extract relevant data
pd = object$pd[ vl & keep, ]
# Dimensions
R = nrow( pd )
C = dim( pd$y )[2]
if ( type == 'p' | type == 'points' ) {
# Extract plotting options
pch = pvt_default_options( pch, pd, ver = 'pch' )
cex = pvt_default_options( cex, pd, ver = 'cex' )
col = pvt_default_options( col, pd, ver = 'col' )
bg = pvt_default_options( bg, pd, ver = 'bg' )
for ( r in 1:R ) {
x = rep( pd$x[r], C )
y = pd$y[r,]
points( x, y, pch = pch[r,], col = col[r,], cex = cex[r,], bg = bg[r,] )
}
}
if ( type == 'l' | type == 'lines' ) {
# Extract plotting options
lty = pvt_default_options( pch, pd, ver = 'lty' )
lwd = pvt_default_options( cex, pd, ver = 'lwd' )
col = pvt_default_options( col, pd, ver = 'col' )
# Loop over test statistic values
for ( c in 1:C ) {
x = pd$x
y = pd$y[,c]
ord = order( x )
lines( x[ord], y[ord], lty = lty[ ord, c ],
lwd = lwd[ ord, c ], col = col[ ord, c ] )
}
}
if ( type == 's' | type == 'segments' ) {
# Extract plotting options
lty = pvt_default_options( pch, pd, ver = 'lty' )
lwd = pvt_default_options( cex, pd, ver = 'lwd' )
col = pvt_default_options( col, pd, ver = 'col' )
for ( u in unique( pd$cv ) ) {
sel = pd$cv == u
for ( c in 1:C ) {
lines( pd$x[sel], pd$y[sel,c],
lty = lty[ sel, c ], lwd = lwd[ sel, c ],
col = col[ sel, c ] )
}
}
}
} else stop( 'No values for a PvT plot', .call = FALSE )
} else {
stop( 'Not an rtplots object', .call = FALSE )
}
}
rettopnivek/rtplots documentation built on May 27, 2019, 5:55 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#----------------------------------#
# Functions for plotting data from #
# simple choice and response time #
# tasks #
#----------------------------------#
# Useful functions for package creation
# library(devtools)
# library(roxygen2)
# * Incomplete documentation
# Index
# Lookup - 01: rtplots*
# Lookup - 02: is.rtplots
# Lookup - 03: lines.rtplots
# Lookup - 04: points.rtplots
# Lookup - 05: plot.rtplots
# Lookup - 06: uncertainty*
# Lookup - 07: draw_ui
# Lookup - 08: add_to_rtplot*
# Lookup - 09
# Lookup - 01
#' Creates a 'rtplots' Object
#'
#' Creates a 'rtplots' object from a data frame of
#' response time and choice/accuracy data, allowing
#' for subsequent plotting.
#'
#' @param df a data frame that includes (at a minimum) a column of
#' response times.
#' @param label a character string giving the column name for
#' the response time variable, and optionally the choice/accuracy
#' variable, a grouping factor variable (e.g., subjects), and
#' finally an additional covariate (e.g., conditions). If
#' \code{NULL}, the algorithm will attempt to locate response
#' time and choice/accuracy variables via a set of default names.
#' @param keep a logical vector, indicating which rows should be
#' kept when extracting the variables.
#' @param type the type of function to compute. Options are
#' \itemize{
#' \item \code{CDF}: the cumulative distribution function.
#' \item \code{QPE}: the quantile function.
#' \item \code{PDF}: the probability density function.
#' \item \code{PVT}: probability versus time.
#' }
#' @param level an optional index indicating the specific level
#' of the grouping factor to consider. This allows, for instance,
#' selecting a single subject.
#' @param ... additional parameters for the density, distribution,
#' and quantile functions.
#'
#' @details
#'
#' Forthcoming
#'
#' @return An object of class 'rtplots', a list consisting of...
#' \itemize{
#' \item \code{pd}; a data frame with the x and y values
#' to plot and their associated choice/accuracy value.
#' \item \code{xm}; If relevant, a matrix whose columns
#' contain the individual x-axis values of the function
#' which were collapsed over the grouping factor.
#' \item \code{ym}; If relevant, a matrix whose columns
#' contain the individual y-axis values of the function
#' which were collapsed over the grouping factor.
#' \item \code{npfd}; a data frame with the number of
#' observations, proportions, and frequencies based on
#' the choice/accuracy values.
#' \item \code{type}; the type of function that was applied.
#' \item \code{x}; the range of x-axis values.
#' \item \code{y}; the range of y-axis values.
#' }
#'
#' @references
#'
#' Forthcoming
#'
#' @examples
#' # Load in example dataset
#' data("priming_data")
#' lbl = c( 'RT', 'Accuracy', 'Subjects' )
#' keep = priming_data$Condition == 0
#' obj = rtplots( priming_data, label = lbl, keep = keep )
#'
#' @export
rtplots = function( df, label = NULL, keep = NULL,
type = 'CDF', level = NULL, ... ) {
# Check input
if ( !is.data.frame( df ) )
stop( 'Input variable must be a data frame' )
output = extract_var( df, label, keep, type )
# Sort RTs into ascending order
output = sort_input( output, level )
# Compute proportions and frequencies
npfd = extract_npfd( output )
if ( check_for_cdf_type(type) ) {
### CDF
# Generate output
tmp = create_cdf_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'CDF'
} else if ( check_for_qpe_type(type) ) {
### QPE
# Generate output
tmp = create_qpe_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'QPE'
} else if ( check_for_pdf_type(type) ) {
### PDF
# Generate output
tmp = create_pdf_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'PDF'
} else if ( check_for_pvt_type(type) ) {
### PVT
# Generate output
tmp = create_pvt_output( output, npfd, ... )
output$pd = tmp$pd
output$xm = tmp$xm
output$ym = tmp$ym
output$npfd = npfd
output$type = 'PVT'
} else {
err = 'That function type is unknown.'
stop( err, call. = FALSE )
}
# Plotting range
output$x = range( na.omit( output$pd$x ) )
output$y = range( output$pd$y )
class( output ) = 'rtplots'
return( output )
}
# Lookup - 02
#' @rdname rtplots
#' @export
is.rtplots = function(x) inherits(x, "rtplots")
# Lookup - 03
#' Lines Method for 'rtplots' Object
#'
#' Provides a method to 'rtplots' objects that allows
#' for drawing line segments on an existing plot.
#'
#' @param object a 'rtplots' object.
#' @param val the choice or accuracy value whose observations
#' are to be plotted.
#' @param ind a index to select observations for a specific
#' choice/accuracy value.
#' @param flip logical; if true, flips the y-axis values to
#' be negative.
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{lines}}.
#'
#' @export
lines.rtplots = function( object, val = NULL,
ind = 1, flip = F, ... ) {
# If it isn't a probability versus time plot
if ( object$type != 'PVT' ) {
if ( !is.null( val ) ) {
vl = val == object$pd$v
} else {
vl = object$val[ ind ] == object$pd$v
}
if ( !flip ) {
lines( object$pd$x[vl], object$pd$y[vl], ... )
} else {
lines( object$pd$x[vl], -object$pd$y[vl], ... )
}
} else { # Special options for PvT figures
# Extract choice value
if ( is.null( val ) ) val = object$val[ ind ]
# Check for multiple RTs per probability
dmn = dim( object$pd$y )
# If the points aren't flipped about the x-axis
if ( !flip ) {
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:dmn[2] ) {
x = object$pd$x[vl]
y = object$pd$y[vl,i]
ord = order(x)
lines( x[ord], y[ord], ... )
}
} else { # Flip
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:dmn[2] ) {
x = object$pd$x[vl]
y = object$pd$y[vl,i]
ord = order(x)
lines( x[ord], -y[ord], ... )
}
}
}
}
# Lookup - 04
#' Points Method for 'rtplots' Object
#'
#' Provides a method to 'rtplots' objects that allows
#' for drawing points on an existing plot.
#'
#' @param object a 'rtplots' object.
#' @param val the choice or accuracy value whose observations
#' are to be plotted.
#' @param ind a index to select observations for a specific
#' choice/accuracy value.
#' @param flip logical; if true, flips the y-axis values to
#' be negative.
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{points}}.
#'
#' @export
points.rtplots = function( object, val = NULL,
ind = 1, flip = F, ... ) {
# If it isn't a probability versus time plot
if ( object$type != 'PVT' ) {
# If a choice value is given
if ( !is.null( val ) ) {
vl = val == object$pd$v
} else {
vl = object$val[ ind ] == object$pd$v
}
# If the points aren't flipped about the x-axis
if ( !flip ) {
points( object$pd$x[vl], object$pd$y[vl], ... )
} else { # Flip
points( object$pd$x[vl], -object$pd$y[vl], ... )
}
} else { # Special options for PvT figures
# Extract choice value
if ( is.null( val ) ) val = object$val[ ind ]
# Check for multiple RTs per probability
dmn = dim( object$pd$y )
# If the points aren't flipped about the x-axis
if ( !flip ) {
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:length(vl) ) {
x = rep( object$pd$x[vl[i]], dmn[2] )
y = object$pd$y[vl[i],]
points( x, y, ... )
}
} else { # Flip
# Subset of values matching desired choice value
vl = which( object$pd$v %in% val )
for ( i in 1:length(vl) ) {
x = rep( object$pd$x[vl[i]], dmn[2] )
y = object$pd$y[vl[i],]
points( x, -y, ... )
}
}
}
}
# Lookup - 05
#' Plot Method for 'rtplots' Object
#'
#' Creates a blank plot based on a 'rtplots' object.
#'
#' @param x a 'rtplots' object.
#' @param y the y-axis boundaries.
#' @param inc the interval for the ticks on the
#' x and y-axes (depending on the plot).
#' @param type if equal to \code{'blank'}, a completely
#' blank plot is generated. Otherwise, converted to \code{'n'}.
#' @param xlab a character string for the label on the x-axis.
#' @param ylab a character string for the label on the y-axis.
#' @param bty the type of box to draw around the plot. See
#' \code{\link[graphics]{par}}.
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{par}}.
#'
#' @export
plot.rtplots = function( x,
y = NULL,
inc = NULL,
type = NULL,
xlab = NULL,
ylab = NULL,
bty = NULL, ... ) {
# Extract 'rtplots' object
object = x
if ( object$type == 'CDF' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
xl = lower_upper( inc[1], object$pd$x )
if ( is.null( y ) ) yl = c(0,1) else yl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Time' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Distribution function'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
if ( object$type == 'PDF' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
xl = lower_upper( inc[1], object$pd$x )
if ( is.null( y ) ) yl = c(0,2) else yl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Time' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Density function'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
if ( object$type == 'QPE' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
xl = lower_upper( inc, na.omit( object$pd$x ) )
if ( is.null( y ) ) yl = c(0,1) else yl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Time' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Quantile function'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
if ( object$type == 'PVT' ) {
# Determine x and y-axis boundaries
if ( is.null( inc ) ) inc = .2
yl = lower_upper( inc, as.vector( na.omit( object$pd$y ) ) )
if ( is.null( y ) ) xl = c(0,1) else xl = y
# Default options for plot
if ( is.null( xlab ) ) xlb = 'Probability' else xlb = xlab
if ( is.null( ylab ) ) ylb = 'Time'
else ylb = ylab
if ( is.null( bty ) ) bx = 'l' else bx = bty
plotYes = T
# Plot type
if ( !is.null( type ) ) {
if ( type == 'blank' | type == 'Blank' ) {
plot( xl, yl, type = 'n', xlab = ' ',
ylab = ' ', xaxt = 'n', yaxt = 'n',
bty = 'n' )
}
plotYes = F
} else type = 'n'
if ( plotYes )
plot( xl, yl, type = type, xlab = xlb, ylab = ylb,
bty = bx, ... )
}
}
# Lookup - 06
#' Computes Uncertainty Intervals for 'rtplots' Objects
#'
#' A function that computes a specified uncertainty
#' interval (such as \eqn{\pm} 2 standard errors from
#' the mean) for a 'rtplots' object.
#'
#' @param object a 'rtplots' object.
#' @param f a function to compute the boundaries of
#' the uncertainty interval given the desired
#' range of coverage (takes at least two parameters,
#' a vector \code{x} and a range \code{interval}).
#' @param alpha the desired width of coverate for the
#' given uncertainty interval.
#' @param ... additional parameters for the
#' function \code{f}.
#'
#' @return Forthcoming
#'
#' @export
uncertainty = function( object, f = NULL,
alpha = .95, ... ) {
# If no function is provided, assume uncertainty
# intervals should be calculated using the
# standard error of the mean
if ( is.null( f ) ) {
f = function(x,interval) {
s = sd( x )/sqrt( length(x) ) # Standard error
m = mean(x)
return( c( m - abs( qnorm(interval[1]) )*s,
m + abs( qnorm(interval[2] ) )*s ) )
}
}
# Determine lower and upper boundaries
interval = numeric(2);
interval[1] = (1 - alpha)/2
interval[2] = interval[1] + alpha
# Initialize output
ud = NULL
pd = NULL
# If there is more than one group
if ( object$n_g > 1 ) {
l = nrow( object$pd )
ud = data.frame( x = rep( NA, l * 3 ),
y = rep( NA, l * 3 ),
loc = rep( NA, l * 3 ),
v = rep( NA, l * 3 ),
i = rep( NA, l * 3 ) )
# Loop over choice/accuracy values
sta = 1; end = 0
for ( i in 1:(object$n_v) ) {
# Compute uncertainty intervals
if ( !is.null( object$xm ) ) {
xui = apply( object$xm[[i]], 2, f, interval = interval,
... )
} else xui = NULL
if ( !is.null( object$ym ) ) {
yui = apply( object$ym[[i]], 2, f, interval = interval,
... )
} else yui = NULL
sel = object$val[ i ] == object$pd$v
# Center values
end = end + sum( sel )
ud$x[ sta:end ] = object$pd$x[ sel ]
ud$y[ sta:end ] = object$pd$y[ sel ]
ud$v[ sta:end ] = object$pd$v[ sel ]
ud$loc[ sta:end ] = 'center'
ud$i[ sta:end ] = 1:sum(sel)
sta = end + 1
# Lower boundary
end = end + sum( sel )
if ( is.null( xui ) )
ud$x[ sta:end ] = object$pd$x[ sel ] else
ud$x[ sta:end ] = xui[1,]
if ( is.null( yui ) )
ud$y[ sta:end ] = object$pd$y[ sel ] else
ud$y[ sta:end ] = yui[1,]
ud$v[ sta:end ] = object$pd$v[ sel ]
ud$loc[ sta:end ] = 'lower'
sta = end + 1
# Upper boundary
end = end + sum( sel )
if ( is.null( xui ) )
ud$x[ sta:end ] = object$pd$x[ sel ] else
ud$x[ sta:end ] = xui[2,]
if ( is.null( yui ) )
ud$y[ sta:end ] = object$pd$y[ sel ] else
ud$y[ sta:end ] = yui[2,]
ud$v[ sta:end ] = object$pd$v[ sel ]
ud$loc[ sta:end ] = 'upper'
sta = end + 1
}
pd = data.frame( xl = rep( NA, l * 2 ),
xu = rep( NA, l * 2 ),
yl = rep( NA, l * 2 ),
yu = rep( NA, l * 2 ) )
pd$xl = c( ud$x[ ud$loc == 'lower' ],
ud$x[ ud$loc == 'center' ] )
pd$xu = c( ud$x[ ud$loc == 'upper' ],
ud$x[ ud$loc == 'center' ] )
pd$yl = c( ud$y[ ud$loc == 'center' ],
ud$y[ ud$loc == 'lower' ] )
pd$yu = c( ud$y[ ud$loc == 'center' ],
ud$y[ ud$loc == 'upper' ] )
pd$v = rep( ud$v[ ud$loc == 'center' ], 2 )
pd$i = rep( ud$i[ ud$loc == 'center' ], 2 )
}
out = list(
pd = pd,
ud = ud,
xm = object$xm,
ym = object$ym,
val = object$val
)
# Define new class
class( out ) = 'rtplots_ui';
return( out )
}
# Lookup - 07
#' Draw Uncertainty Intervals
#'
#' A convenience function that draws uncertainty
#' intervals computed from a 'rtplots' object onto
#' an already existing plot.
#'
#' @param ui a 'rtplots_ui' object.
#' @param val the choice/accuracy value for which to
#' draw the uncertainty intervals.
#' @param ind the index of the desired choice/accuracy
#' value over which to draw uncertainty intervals.
#' @param flip logical; if \code{true} the y-axis
#' values are flipped to be negative.
#' @param type the type of method used to draw the
#' uncertainty interval. Options are
#' \itemize{
#' \item \code{segments}.
#' \item \code{arrows}.
#' \item \code{polygon}.
#' }
#' @param ... additional plotting parameters. See
#' \code{\link[graphics]{segments}},
#' \code{\link[graphics]{arrows}}, and
#' \code{\link[graphics]{polygon}}.
#'
#' @export
draw_ui = function( ui, val = NULL, ind = 1,
flip = F, type = 'segments', ... ) {
if ( !is.null( val ) ) {
vl = val == ui$pd$v
} else {
vl = ui$val[ ind ] == ui$pd$v
}
if (!flip) w = 1 else w = -1
segment_names = c( 'segments', 's', 'seg' )
if ( type %in% segment_names ) {
segments( ui$pd$xl[vl], w*ui$pd$yl[vl],
ui$pd$xu[vl], w*ui$pd$yu[vl], ... )
}
polygon_names = c( 'polygon', 'p', 'poly' )
if ( type %in% polygon_names ) {
polygon( c( ui$pd$xl[vl], rev( ui$pd$xu[vl] ) ),
c( ui$pd$yu[vl], rev( ui$pd$yl[vl] ) ),
... )
}
arrow_names = c( 'arrows', 'a', 'arr' )
if ( type %in% arrow_names ) {
arrows( ui$pd$xl[vl], w*ui$pd$yl[vl],
ui$pd$xu[vl], w*ui$pd$yu[vl], ... )
}
}
# Lookup - 08
#' Add Points or Lines to a RT Plot
#'
#' A convenience function that draws points or
#' line segments onto an already existing plot. Useful
#' for adding summary statistics (e.g., the mean or
#' median).
#'
#' @param object a rtplots object or a list with the
#' x and y-axis values to plot.
#' @param T_x the test statistic(s) to be added to the
#' existing figure.
#' @param T_g the test statistic to use when collapsing
#' over a grouping factor.
#' @param val Forthcoming.
#' @param ind Forthcoming.
#' @param type Forthcoming.
#' @param out Forthcoming.
#' @param ... Forthcoming.
#'
#' @export
add_to_rtplot = function( object, T_x = mean, T_g = mean, val = NULL,
ind = 1, type = 'p', out = T, ... ) {
# Function to calculate values for y-axis
# using linear interpolation
f = function( x ) {
ind = max( which( xa < x ) )
if ( ind == -Inf ) {
out = ya[1];
} else {
if ( ind == length(xa) ) {
out = ya[ length(xa) ]
} else {
pts = c( ind, ind + 1 )
out = ln_interp( x, ya[ pts ], xa[ pts ] )
}
}
return( out )
}
if ( is.rtplots( object ) ) {
# If it isn't a probability versus time plot
if ( object$type != 'PVT' ) {
# If there is no grouping factor
if ( object$n_g <= 1 ) {
if ( !is.null( val ) ) {
vl = val == object$ad$v
vly = val == object$pd$v
} else {
vl = object$val[ ind ] == object$ad$v
vly = object$val[ ind ] == object$pd$v
}
ts = T_x( object$ad$t[vl] )
# Extract plotting values
xa = object$pd$x[vly] # x-axis
ya = object$pd$y[vly] # y-axis
# Determine y-axis values
ny = sapply( ts, f )
output = list( x = ts, y = ny )
} else {
if ( !is.null( val ) ) {
vl = val == object$ad$v
vly = val == object$pd$v
} else {
vl = object$val[ ind ] == object$ad$v
vly = object$val[ ind ] == object$pd$v
}
dmn = length( T_x( object$ad$t[vl] ) )
ts_all = by( object$ad$t[vl], list( object$ad$g[vl] ), T_x )
ts_all = matrix( unlist( ts_all ), length( ts_all ), dmn, byrow = T )
ts = apply( ts_all, 2, T_g )
# Extract plotting values
xa = object$pd$x[vly] # x-axis
ya = object$pd$y[vly] # y-axis
# Determine y-axis values
ny = sapply( ts, f )
output = list( x = ts, y = ny )
}
}
} else {
pass = F
if ( is.list( object ) ) {
if ( all( c('x','y') %in% names( object ) ) ) {
pass = T
output = object
}
if (!pass) stop( 'Please pass in an "rtplots" object',
.call = FALSE )
}
}
if ( type == 's' | type == 'segments' ) {
segments( output$x, 0, output$x, output$y, ... )
}
if ( type == 'p' | type == 'points' ) {
points( output$x, output$y, ... )
}
if ( out ) return( output )
}
# Lookup - 09
#' Create a Probability versus Time Figure
#'
#' A convenience function that draws points, lines, or
#' line segments for a probability versus time figure
#' onto an already existing plot. Allows fine-tune
#' control of plotting parameters such as point type,
#' size, etc.
#'
#' @param object a rtplots object.
#'
#'
#' @export
pvt_plot = function( object, keep = NULL, ind = NULL, val = NULL, type = 'p',
pch = NULL, cex = NULL, col = NULL,
bg = NULL, lwd = NULL ) {
if ( is.rtplots( object ) ) {
if ( object$type == 'PVT' ) {
# If a choice value is given
if ( !is.null( val ) ) {
vl = val == object$pd$v
} else if ( !is.null( ind ) ) {
vl = object$val[ ind ] == object$pd$v
} else {
vl = rep( T, nrow( object$pd ) )
}
# Extract relevant data
if ( is.null( keep ) ) keep = rep( T, nrow( object$pd ) )
# Extract relevant data
pd = object$pd[ vl & keep, ]
# Dimensions
R = nrow( pd )
C = dim( pd$y )[2]
if ( type == 'p' | type == 'points' ) {
# Extract plotting options
pch = pvt_default_options( pch, pd, ver = 'pch' )
cex = pvt_default_options( cex, pd, ver = 'cex' )
col = pvt_default_options( col, pd, ver = 'col' )
bg = pvt_default_options( bg, pd, ver = 'bg' )
for ( r in 1:R ) {
x = rep( pd$x[r], C )
y = pd$y[r,]
points( x, y, pch = pch[r,], col = col[r,], cex = cex[r,], bg = bg[r,] )
}
}
if ( type == 'l' | type == 'lines' ) {
# Extract plotting options
lty = pvt_default_options( pch, pd, ver = 'lty' )
lwd = pvt_default_options( cex, pd, ver = 'lwd' )
col = pvt_default_options( col, pd, ver = 'col' )
# Loop over test statistic values
for ( c in 1:C ) {
x = pd$x
y = pd$y[,c]
ord = order( x )
lines( x[ord], y[ord], lty = lty[ ord, c ],
lwd = lwd[ ord, c ], col = col[ ord, c ] )
}
}
if ( type == 's' | type == 'segments' ) {
# Extract plotting options
lty = pvt_default_options( pch, pd, ver = 'lty' )
lwd = pvt_default_options( cex, pd, ver = 'lwd' )
col = pvt_default_options( col, pd, ver = 'col' )
for ( u in unique( pd$cv ) ) {
sel = pd$cv == u
for ( c in 1:C ) {
lines( pd$x[sel], pd$y[sel,c],
lty = lty[ sel, c ], lwd = lwd[ sel, c ],
col = col[ sel, c ] )
}
}
}
} else stop( 'No values for a PvT plot', .call = FALSE )
} else {
stop( 'Not an rtplots object', .call = FALSE )
}
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.