R/02_data_preparation.R
In mrelnoob/jk.dusz.tarping: Prepare and Analyse the Tarping Survey Data
Defines functions
cond.boxplots
uni.histograms
uni.simudistrib
uni.dotplots
uni.boxplots
model_datasets
Documented in
cond.boxplots
model_datasets
uni.boxplots
uni.dotplots
uni.histograms
uni.simudistrib
############################
##### DATA PREPARATION #####
############################
### IMPORTANT NOTE: to avoid creating notes for unquoted variables, I must add the following code at
# the beginning of every source file (e.g. R/myscript.R) that uses unquoted variables, so in front of
# all my scripts doing any kind of analyses (otherwise, I should always assign each variable, e.g.
# mydata$myvariable, which is quite time consuming and wearisome).
if(getRversion() >= "2.15.1") utils::globalVariables(c(
"manager_id", "xp_id", "country", "latitude", "longitude", "elevation", "tarping_date",
"planned_duration", "goals",
"restoration", "operation_type", "multiple_ops", "freq_monitoring", "slope", "difficulty_access",
"shade", "forest", "ruggedness", "granulometry", "obstacles", "flood", "environment", "fabric_type",
"liner_geomem", "agri_geomem", "woven_geotex", "mulching_geotex", "pla_geotex", "weedsp_geotex",
"other_unknown", "grammage", "thickness", "resi_punc", "resi_trac", "season", "maxveg", "preparation",
"levelling", "stand_surface", "age", "fully_tarped", "distance", "multi_strips", "strips_overlap",
"stripsoverlap_ok", "tarpfix_pierced", "geomem", "geotex", "uprootexcav", "efficiency", "high_eff",
"strips_fixation", "staples_distance", "fabric_fixation", "tarpfix_multimethod", "sedicover_height",
"trench", "trench_depth", "pierced_tarpinstall", "plantation", "repairs", "add_control",
"add_control_type", "degradation", "pb_fixation", "pb_durability", "pb_trampiercing", "pb_vandalism",
"regrowth_during", "reg_staples", "reg_stripsoverlap", "reg_obstacles", "reg_holes", "reg_plantations",
"reg_pierced", "reg_edges", "reg_nearby", "untarped_regrowth", "tarping_abandoned", "tarping_completed",
"tarping_ongoing", "tarping_duration", "latest_condition", "latest_regrowth", "latest_months",
"eff_expansion", "eff_dispersal", "eff_vigour", "eff_eradication"))
# Alternatively, here's another solution (to insert within a given function) when the number of unquoted
# variables is low:
# planned_duration <- age <- plantation <- NULL
# Avoids potentially problematic NOTE during the R CMD check (i.e. devtools::check()) due to the
# fact that these 3 variables are un-quoted (also known as non-standard evaluation (NSE)). If you want
# to find again how I know that, just run a Google search with some of the NOTE message "package: no
# visible binding for global variable" (I read a r-blogger.com post about it).
# _________________________________________
### Creation of a function that will create separate datasets for each response variable (depending on the
# explanatory variables their models will potentially include):
#' Datasets for Models Building
#'
#' @description The `model_datasets` function creates the respective reduced datasets that should be used
#' to model each response variables. For instance, if `response.var = "overlaps"`, the function will
#' produce a dataset containing \emph{reg_stripsoverlap} as the \strong{response} variable and a subset of
#' variables to be used as \strong{predictors/explanatory variables} (removing all the variables that
#' should not be used to model \emph{reg_stripsoverlap}'s variations).
#' @param response.var A character string specifying which modelling dataset should be produced (either:
#' efficiency", "edges", "overlaps", or "tarpedarea"):
#' * "efficiency" will produce the dataset for the 3 efficiency evaluation variables (namely \emph{
#' eff_eradication}, \emph{eff_expansion} and \emph{eff_vigour});
#' * "edges" will produce the dataset having for response variable the tarping operations that observed
#' regrowth at the edge of the tarped area (at the end of the operation or during the latest visit to the
#' site);
#' * "overlaps" produce the dataset having for response variable the tarping operations that observed
#' regrowth at strip overlaps;
#' * "tarpedarea" will produce the dataset having for response variable the tarping operations that observed
#' regrowth at somewhere on the area covered by the fabric.
#'
#' @return A tibble.
#' @export
#' @import dplyr stringr
#'
#' @examples
#' \dontrun{
#' eff_model <- model_datasets(response.var = "efficiency")
#' }
model_datasets <- function(response.var = c("efficiency", "edges", "overlaps",
"tarpedarea")){
ppp <- jk.dusz.tarping::clean_my_data()
ppp %>%
dplyr::filter(!stringr::str_detect(operation_type, "crushing_tarping_trial")) %>%
dplyr::mutate(uprootexcav = ifelse(preparation == "excavation" |
preparation == "uprooting", 1, 0)) -> qqq # Exclude rows which
# belong to crushing-tarping trials, and creates a new variable called "uprootexcav"!
### For the 3 "efficiency" models
if (response.var == "efficiency") {
qqq <- dplyr::mutate(.data = qqq,
high_eff = ifelse(stringr::str_detect(latest_regrowth, stringr::regex("few.", dotall = TRUE)) |
eff_eradication == "1", 1, 0)) %>%
dplyr::mutate_if(is_binary, factor)
# With str_detect and regex, I matched all obs. with latest_regrowth that begins with "few"
# (quite similar to grep)!
qqq <- dplyr::mutate(.data = qqq,
efficiency = rowMeans(x = qqq[,80:82], na.rm = FALSE))
tapioca <- qqq[,c("manager_id", "xp_id", "efficiency", "eff_eradication", "high_eff",
"latitude", "longitude", "elevation", "goals",
"freq_monitoring", "slope", "difficulty_access", "shade", "forest", "ruggedness",
"granulometry", "obstacles", "flood",
"geomem", "geotex", "liner_geomem", "agri_geomem", "woven_geotex", "mulching_geotex",
"pla_geotex", "weedsp_geotex", "other_unknown", "grammage", "thickness",
"maxveg", "uprootexcav", "stand_surface", "age", "fully_tarped", "distance", "tarping_duration",
"stripsoverlap_ok", "tarpfix_multimethod", "tarpfix_pierced", "sedicover_height", "trench_depth",
"pierced_tarpinstall", "plantation", "repairs", "add_control", "add_control_type",
"degradation", "pb_fixation", "pb_durability")]
}
### For the "lreg_edges" model
if (response.var == "edges") {
qqq <- dplyr::mutate(.data = qqq, lreg_edges =
ifelse(c(grepl("*edges*", latest_regrowth) | grepl("*edges*", untarped_regrowth)), 1, 0)) %>%
dplyr::mutate(.data = qqq, reg_elsewhere =
ifelse(reg_staples == 1 | reg_stripsoverlap == 1 | reg_obstacles == 1 |
reg_holes == 1 | reg_plantations == 1 | reg_pierced == 1, yes = 1, no = 0)) %>%
dplyr::filter(fully_tarped == 1) # Only for fully tarped operations!
tapioca <- qqq[,c("manager_id", "xp_id", "lreg_edges",
"latitude", "longitude", "elevation", "slope", "flood", "difficulty_access", "shade", "forest",
"ruggedness", "granulometry", "obstacles", "stand_surface",
"geomem", "geotex",
"maxveg", "uprootexcav",
"distance", "stripsoverlap_ok",
"tarpfix_multimethod", "tarpfix_pierced", "sedicover_height", "trench_depth", "plantation",
"repairs", "add_control", "freq_monitoring",
"degradation", "pb_fixation", "pb_durability", "tarping_duration", "reg_elsewhere")]
}
### For the "overlaps" model
if (response.var == "overlaps") {
qqq <- dplyr::mutate(.data = qqq, reg_elsewhere =
ifelse(reg_staples == 1 | reg_edges == 1 | reg_obstacles == 1 |
reg_holes == 1 | reg_plantations == 1 | reg_pierced == 1, yes = 1, no = 0)) %>%
dplyr::mutate(stripfix_pierced =
ifelse(c(grepl("*staples*", strips_fixation) | grepl("*stakes*", strips_fixation)), 1, 0)) %>%
dplyr::mutate(stripfix_taped =
ifelse(c(strips_fixation == "tape_and_heavyobj" | strips_fixation == "double_tape" |
strips_fixation == "tape" | strips_fixation == "tape_and_staples"), 1, 0)) %>%
dplyr::filter(multi_strips == 1) %>%
dplyr::mutate_if(is_binary, factor)
tapioca <- qqq[,c("manager_id", "xp_id", "reg_stripsoverlap",
"latitude", "longitude", "elevation", "slope", "flood", "difficulty_access","shade", "forest", "ruggedness",
"granulometry", "obstacles", "stand_surface",
"geomem", "geotex",
"maxveg", "uprootexcav", "levelling", "fully_tarped", "distance",
"strips_overlap", "tarpfix_multimethod", "tarpfix_pierced", "stripfix_pierced", "stripfix_taped",
"sedicover_height", "trench_depth", "plantation",
"repairs", "add_control", "freq_monitoring", "pb_fixation", "pb_durability",
"reg_elsewhere", "tarping_duration")]
}
### For the "tarpedarea" model
if (response.var == "tarpedarea") {
qqq <- dplyr::mutate(.data = qqq, lreg_tarpedarea =
ifelse(c(grepl("*tarpedarea*", latest_regrowth) |
grepl("*tarpedarea*", untarped_regrowth)), 1, 0))
tapioca <- qqq[,c("manager_id", "xp_id", "lreg_tarpedarea",
"latitude", "longitude", "elevation", "slope", "flood",
"difficulty_access", "shade", "forest", "ruggedness", "granulometry", "obstacles", "stand_surface",
"geomem", "geotex", "woven_geotex",
"maxveg", "uprootexcav", "levelling", "fully_tarped", "distance",
"stripsoverlap_ok", "tarpfix_multimethod", "tarpfix_pierced", "pierced_tarpinstall",
"sedicover_height", "trench_depth", "plantation",
"repairs", "add_control", "freq_monitoring",
"pb_fixation", "pb_durability", "pb_trampiercing", "reg_edges", "tarping_duration")]
}
return(tapioca)
}
# _________________________________________
### Creation of a function that draws univariate boxplots for all numeric variables in a given dataset:
#' Univariate boxplots
#'
#' @description The `uni.boxplots` function draws, within a single panel, an independent boxplot for each
#' numeric (continuous or discrete) variable in a given dataset. It is particularly useful for data
#' exploration (e.g. Zuur \emph{et al.}, 2010). For instance, to simultaneously observe the
#' distributions of all numeric variables or \strong{to detect their univariate outliers}.
#'
#' @details The `uni.boxplots` function only modifies the graphical parameters of the
#' \code{\link[graphics:boxplot]{boxplot}} function in the `graphics` package to match some predefined
#' preferences. Therefore, default values of `uni.boxplots` create nice looking boxplots but retain
#' default \emph{heuristic} aspects of `boxplot` (such as the length of whiskers or the plotting of
#' outliers). These aspects can however be changed as in \code{\link[graphics:boxplot]{boxplot}}. \cr
#' On the other hand, panel parameters are internally controlled using `par`. However, to avoid unforeseen
#' conflicts with other internal parameters, it is not possible to tune panel parameters as we would
#' do with `par`. Instead, parametrization is only possible with the given subset of parameters.
#'
#' @note To avoid \emph{recursive argument errors}, internal arguments should be called using upper case
#' letters (e.g. CEX.LAB = 0.9) whereas other arguments from the `boxplot` function should be called with
#' their normal case writing (e.g. outline = FALSE)!
#'
#' @param dataset The input dataset containing all variables to be plotted. It may contain all kinds of
#' variables, the `uni.boxplots` function will automatically detect and plot numeric variables (columns).
#' @param ... Any other parameter that can be incorporated in \code{\link[graphics:boxplot]{boxplot}}.
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(0.5,4.1,1.1,1.5)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting of `cex`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(2.4, 0.6, 0).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param TYPE The type of boxplot to draw. Default is "n".
#' @param BORDER An optional vector of colors for the outlines of the boxplots. The values in border
#' are recycled if the length of border is less than the number of plots. Default is "lightcoral".
#' @param COL If col is non-null it is assumed to contain colors to be used to colour the bodies of
#' the boxplots. Default is "moccasin".
#' @param LTY The line type. Line types can either be specified as an integer (0=blank, 1=solid (default),
#' 2=dashed, 3=dotted, 4=dotdash, 5=longdash, 6=twodash) or as one of the character strings "blank",
#' "solid", "dashed", "dotted", "dotdash", "longdash", or "twodash", where "blank" uses ‘invisible lines’
#' (i.e., does not draw them).
#' @param STAPLEWEX Staple line width expansion, proportional to box width. Default is 0.
#' @param WHISKLWD Whisker line width expansion. Default is 2.
#' @param BOXWEX A scale factor to be applied to all boxes. When there are only a few groups, the
#' appearance of the plot can be improved by making the boxes narrower. Default is 0.7.
#' @param BOXLWD Width of boxplot outer lines. Default is 0.1.
#' @param MEDLWD Width of the median line. Default is 2.6.
#' @param PCH The type of points to be drawn for outliers. Default is 19. See \code{\link[graphics:points]{points}}
#' for possible values and their interpretation.
#'
#' @return A panel of univariate boxplots.
#' @export
#' @import graphics
#'
#' @examples
#' data("mtcars")
#' uni.boxplots(dataset = mtcars)
uni.boxplots <- function(dataset, MAR=c(0.5,4.1,1.1,1.5), CEX.LAB=1, FONT.LAB=2, BTY = "n", FG = "gray35",
COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.8, TCL = -0.3,
MGP = c(2.4, 0.6, 0), OMA = c(1, 0, 0, 0),
TYPE = "n", BORDER = "lightcoral", COL = "moccasin", LTY = 1, STAPLEWEX = 0,
WHISKLWD = 2, BOXWEX = 0.7, BOXLWD = 0.1, MEDLWD = 2.6, PCH = 19, ...){
num.data <- dataset[, sapply(dataset, is.numeric)]
nam <- names(num.data)
ncol.data <- ncol(num.data)
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
graphics::par(mfrow= c(ncol.adjust,4), mar=MAR, cex.lab=CEX.LAB, font.lab=FONT.LAB, bty=BTY, fg=FG,
col.axis=COL.AXIS, col.lab=COL.LAB, cex=CEX.PAR, tcl=TCL, mgp=MGP, oma=OMA)
for (i in c(1:ncol.data)) {
graphics::boxplot(num.data[,i],ylab =(nam[i]), type=TYPE, border=BORDER, col = COL,
lty=LTY, staplewex=STAPLEWEX, whisklwd=WHISKLWD, boxwex=BOXWEX, boxlwd=BOXLWD,
medlwd=MEDLWD, pch=PCH, cex=0.7, ...) }
}
# _________________________________________
### Creation of a function that draws univariate Cleveland dotplots for all numeric variables in a
### given dataset:
#' Univariate Cleveland dotplots
#'
#' @description The `uni.dotplots` function draws, within a single panel, an independent Cleveland dotplot
#' (i.e. plotting value against rank) for each numeric (continuous or discrete) variable in a given
#' dataset. It is particularly useful for data exploration (e.g. Zuur \emph{et al.}, 2010). For instance,
#' to simultaneously observe the distributions of all numeric variables or \strong{to detect their
#' univariate outliers}.
#'
#' @details The `uni.dotplots` function only modifies the graphical parameters of the
#' \code{\link[graphics:plot.default]{plot}} function in the `graphics` package to match some predefined
#' preferences. Therefore, default values of `uni.dotplots` create nice looking dotplots but retain
#' some aspects of the original `plot` function. These aspects can however be changed as in
#' \code{\link[graphics:plot.default]{plot}}. \cr
#' On the other hand, panel parameters are internally controlled using `par`. However, to avoid unforeseen
#' conflicts with other internal parameters, it is not possible to tune panel parameters as we would
#' do with `par`. Instead, parametrization is only possible with the given subset of parameters.
#'
#' @note To avoid \emph{recursive argument errors}, internal arguments should be called using upper case
#' letters (e.g. CEX.LAB = 0.9) whereas other arguments from the `plot` function should be called with
#' their normal case writing (e.g. sub = "My subtitle")!
#'
#' @param dataset The input dataset containing all variables to be plotted (must be a `data.frame` with
#' at least 2 variables). It may contain all kinds of columns, the `uni.dotplots` function will
#' automatically detect and plot numeric variables (columns).
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(0.5,4.1,1.1,1.5)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting
#' of `CEX.PAR`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(2.4, 0.6, 0).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param LAB A numerical vector of the form `c(x, y, len)` which modifies the default way that axes
#' are annotated. The values of `x` and `y` give the (approximate) number of tickmarks on the x and y
#' axes and len specifies the label length. The default is `c(5, 5, 7)`. Note that this only affects
#' the way the parameters xaxp and yaxp are set when the user coordinate system is set up, and is
#' not consulted when axes are drawn. `len` is unimplemented in `R`.
#' @param COL.PCH The color to be used for points. Default is "lightcoral".
#' @param PCH The type of points to be drawn. Default is 19. See \code{\link[graphics:points]{points}}
#' for possible values and their interpretation.
#' @param COL.GRID The color of the background grid. Default is "lavender".
#' @param NX The number of lines for the grid in x. Default value is 5.
#' @param NY The number of lines for the grid in Y. Default value is 9.
#' @param LTY The type of lines to be drawn in the background grid. Default value is 6.
#' @param ... Any other parameter that can be incorporated in \code{\link[graphics:plot.default]{plot}}.
#'
#' @return A panel of univariate dotplots.
#' @export
#' @import graphics
#'
#' @examples
#' data("mtcars")
#' uni.dotplots(dataset = mtcars, COL.GRID = "lightblue", LTY = 1, NX = 10, NY = 20)
uni.dotplots <- function(dataset, MAR=c(3,2,0.5,1.5), CEX.LAB = 1.2, FONT.LAB = 2, BTY = "n",
FG = "gray35", COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.6,
TCL = -0.3, MGP = c(1.7, 0.6, 0.1), OMA = c(1, 0, 1, 0), LAB = c(5, 10, 7),
COL.PCH = "lightcoral", PCH = 19, COL.GRID = "lavender", NX = 5, NY = 9, LTY = 6,
...){
num.data <- dataset[, sapply(dataset, is.numeric)]
nam <- names(num.data)
ncol.data <- ncol(num.data)
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
num.data <- as.matrix(num.data)
graphics::par(mfrow= c (ncol.adjust,4), mar=MAR, cex.lab = CEX.LAB, font.lab=FONT.LAB, bty = BTY, fg = FG,
col.axis = COL.AXIS, col.lab = COL.LAB, cex = CEX.PAR, tcl = TCL,
mgp = MGP, oma = OMA, lab = LAB)
for (i in c(1:ncol(num.data))) {
graphics::plot(x = num.data[,i], y = 1:length(num.data[,i]), type = "p", xlab = nam[i], ylab = "",
col = COL.PCH, pch = PCH, panel.first = {
grid(col=COL.GRID,nx = NX,ny = NY, lty = LTY)
}, ...) }
# Here, the argument panel.first={} is used to draw the grid first, so behind the points!
}
# _________________________________________
### Creation of a function that generate random samples from various distributions based on my variables
### parameters:
#' Random sample simulation
#'
#' @description The `uni.simudistrib` function **automatically generates 5 Cleveland dotplots** of random
#' samples from different distributions (either `normal`, `log-normal`, or `poisson`) based on the
#' parameters of the variables in the input `data.frame` or `matrix` (see \emph{Details}). \cr
#' This function is useful to see whether variables' extreme values are actual outliers or whether they
#' lie in a range of values possible for a random sample drawn from a `normal`, `log-normal`, or a `poisson`
#' distribution. \emph{In fine}, it may help determine if the original variable can be approximated by
#' these distribution with or without a transformation.
#'
#' @details The `uni.simudistrib` function extracts some key parameters from the input variables (sample
#' size, mean and standard deviation) and generates random samples based on these parameters. For instance,
#' if `simu.var` contains \emph{i} variables `X1`, `X2`, ... `Xi` and if `distribution = "normal"`, the
#' function will return a panel of \emph{i}x5 plots:
#'
#' * The 1st row will contain five dotplots for five random samples with \emph{n} = `length(X1)` and drawn
#' from a Normal distribution with the same mean and standard deviation as `X1`.
#' * The 2nd row will contain five dotplots for five random samples with \emph{n} = `length(X2)` and drawn
#' from a Normal distribution with the same mean and standard deviation as `X2`.
#' * Etc.
#'
#' \strong{Warning}: the function may fail for `log-normal` and `poisson` distributions if input
#' variables contain negative values (because these distributions are by definition positive). Additionally,
#' if `distribution = "poisson"`, the resulting plots will return \strong{integer} values as Poisson
#' is a discrete probability distribution.
#'
#' @param simu.var A `data.frame` or a `matrix`. For obvious layout and readability reasons, `simu.var`
#' should not include too many variables (`p` < 12 is advised) if the plot is to be printed in a page
#' or window of limited dimensions. For any use in HTML documents (e.g. with `RMarkdown`), the number of
#' input variables should not be a problem.
#' @param distribution Either `"normal"`, `"log-normal"` or `"poisson"` (case sensitive).
#'
#' @return A panel of `p`*5 plots, where `p` is the number of variables in `simu.var`.
#' @export
#' @import graphics
#' @importFrom stats na.omit sd rnorm rlnorm rpois
#'
#' @examples
#' uni.simudistrib(simu.var = iris[,1:4], distribution = "normal")
uni.simudistrib <- function(simu.var, distribution){
if (!is.data.frame(simu.var) && !is.matrix(simu.var)) {
warning("The input dataset (i.e. simu.var) must either be a data.frame or a matrix!")
}
if(missing(distribution)){
stop("Please specify the desired distribution. See ?uni.simudistrib for details.")
}
ncol.data <- ncol(simu.var)
nam <- names(simu.var)
num.data <- as.matrix(simu.var)
graphics::par(mfrow = c(ncol.data,5), mar = c(3,2,0.1,1.5), cex.lab = 1, font.lab=2, bty = "n",
fg = "gray35", col.axis = "gray35", col.lab = "gray20", cex = 0.6, tcl = -0.3,
mgp = c(1.8, 0.6, 0.1), oma = c(0.2, 0.2, 0, 0), lab = c(5, 10, 7))
for (i in c(1:ncol(num.data))) {
mu <- mean(stats::na.omit(num.data[,i]))
std <- stats::sd(stats::na.omit(num.data[,i]))
n <- length(num.data[,i])
rrr <- NULL
if (distribution == "normal") {
for (j in 1:5) {
rrr[[j]] <- stats::rnorm(n = n, mean = mu, sd = std)
graphics::plot(x = rrr[[j]], y = 1:length(rrr[[j]]), type = "p",
xlab = paste("rnorm based on", nam[i], sep = " "), ylab = "", col = "hotpink3",
pch = 19, panel.first = {grid(col="lavender",nx = 9,ny = 3, lty = 6)})}
}
if (distribution == "log-normal") {
for (j in 1:5) {
rrr[[j]] <- stats::rlnorm(n = n,
meanlog = log(mu^2 / sqrt(std^2 + mu^2)),
sdlog = sqrt(log(1 + (std^2 / mu^2)))) # That's how you use rlnorm!!!!!!
graphics::plot(x = rrr[[j]], y = 1:length(rrr[[j]]), type = "p",
xlab = paste("rlnorm based on", nam[i], sep = " "), ylab = "", col = "hotpink3",
pch = 19, panel.first = {grid(col="lavender",nx = 9,ny = 3, lty = 6)})}
}
if (distribution == "poisson") {
for (j in 1:5) {
rrr[[j]] <- stats::rpois(n = n, lambda = mu)
graphics::plot(x = rrr[[j]], y = 1:length(rrr[[j]]), type = "p",
xlab = paste("rpois based on", nam[i], sep = " "), ylab = "", col = "hotpink3",
pch = 19, panel.first = {grid(col="lavender",nx = 9,ny = 3, lty = 6)})}
}
}
}
# _________________________________________
### Creation of a function that generates histograms for each numeric variable:
#' Histogram Panel
#'
#' @description The `uni.histograms` function draws, within a single panel, an independent histogram
#' or each numeric (continuous or discrete) variable in a given dataset. It is particularly useful
#' for data exploration (e.g. Zuur \emph{et al.}, 2010). For instance, to simultaneously observe
#' the distributions of all numeric variables and determine which one will require transformation.
#'
#' @param dataset The input dataset containing all variables to be plotted (must be a `data.frame` with
#' at least 2 variables). It may contain all kinds of columns, the `uni.histograms` function will
#' automatically detect and plot numeric variables (columns).
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(0.5,4.1,1.1,1.5)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting
#' of `CEX.PAR`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(2.4, 0.6, 0).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param LAB A numerical vector of the form `c(x, y, len)` which modifies the default way that axes
#' are annotated. The values of `x` and `y` give the (approximate) number of tickmarks on the x and y
#' axes and len specifies the label length. The default is `c(5, 5, 7)`. Note that this only affects
#' the way the parameters xaxp and yaxp are set when the user coordinate system is set up, and is
#' not consulted when axes are drawn. `len` is unimplemented in `R`.
#' @param BREAKS One of:
#' * a vector giving the breakpoints between histogram cells,
#' * a function to compute the vector of breakpoints,
#' * a single number giving the number of cells for the histogram,
#' * a character string naming an algorithm to compute the number of cells (see
#' \code{\link[graphics:hist]{hist}}),
#' * a function to compute the number of cells.
#'
#' In the last three cases the number is a suggestion only; as the breakpoints will be set to pretty
#' values, the number is limited to 1e6 (with a warning if it was larger). If breaks is a function,
#' the x vector is supplied to it as the only argument (and the number of breaks is only limited by
#' the amount of available memory).
#' @param COL The color of the bar of the histograms (bins).
#' @param BORDER The color of the border of the bars.
#'
#' @return A panel of histograms.
#' @import graphics
#' @export
#'
#' @examples
#' uni.histograms(dataset = iris[,1:4])
uni.histograms <- function(dataset, MAR=c(3,2,0.5,1.5), CEX.LAB = 1.2, FONT.LAB = 2, BTY = "n",
FG = "gray35", COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.6,
TCL = -0.3, MGP = c(1.7, 0.6, 0.1), OMA = c(1, 0, 1, 0), LAB = c(5, 10, 7),
BREAKS = 10, COL = "moccasin", BORDER = "white"){
num.data <- dataset[, sapply(dataset, is.numeric)]
nam <- names(num.data)
ncol.data <- ncol(num.data)
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
num.data <- as.matrix(num.data)
graphics::par(mfrow= c (ncol.adjust,4), mar=MAR, cex.lab = CEX.LAB, font.lab=FONT.LAB, bty = BTY, fg = FG,
col.axis = COL.AXIS, col.lab = COL.LAB, cex = CEX.PAR, tcl = TCL,
mgp = MGP, oma = OMA, lab = LAB)
for (i in c(1:ncol(num.data))) {
graphics::hist(num.data[,i], breaks = BREAKS, col = COL, border = BORDER,
main = "", xlab = nam[i], ylab = "")
}
}
# ___________________________________________________________
### Creation of a function that generates conditional boxplots between the given variables:
#' Bivariate Conditional Boxplots
#'
#' @description The `cond.boxplots` function draws, within a single panel, conditional boxplots (i.e.
#' a plot of a continuous Y variable against one or several categorical X(s) variable(s) (factor),
#' or \strong{discretized} numeric variable(s)) in order to explore bivariate relationships and
#' assess \strong{heteroscedasticity}. \cr
#' If some of the X variables are quantitative (numeric), they will consequently be transformed into
#' factors and discretized into a given number of classes, set approximately by the `N` parameter. This
#' factorisation is required because boxplots are not meant to plot two quantitative variables (without
#' factorisation, the function would plot as many boxplots as there are values in X).
#'
#' @param dataset The input `data.frame` containing the variables to plot (both the Y and the X(s)). Must
#' only contain numeric or factor variables!
#' @param Y The number of the column of `dataset` containing the variable to be plotted as Y. This
#' parameter should be specified as an integer. For instance, if the variable to be used as Y is called
#' "blip" and is the second column of the `dataset`, the `Y` parameter should be `2` (and NOT "two",
#' `dataset$blip`, or `dataset[,2]`).
#' @param Xs The numbers of the columns to be plotted as Xs (e.g. `4:10`, if all the columns from the
#' fourth to the tenth are to be plotted against `Y`).
#' @param outlier Logical (`TRUE` or `FALSE`). Should outliers be plotted or not?
#' @param N Integer. The number of bins on X. Default is 6.
#'
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(2.2,2.1,0.5,1.7)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting
#' of `CEX.PAR`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(1.2, 0.4, 0.2).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param TYPE The type of boxplot to draw. Default is "n".
#' @param BORDER An optional vector of colors for the outlines of the boxplots. The values in border
#' are recycled if the length of border is less than the number of plots. Default is "lightcoral".
#' @param COL If col is non-null it is assumed to contain colors to be used to colour the bodies of
#' the boxplots. Default is "moccasin".
#' @param LTY The line type. Line types can either be specified as an integer (0=blank, 1=solid (default),
#' 2=dashed, 3=dotted, 4=dotdash, 5=longdash, 6=twodash) or as one of the character strings "blank",
#' "solid", "dashed", "dotted", "dotdash", "longdash", or "twodash", where "blank" uses ‘invisible lines’
#' (i.e., does not draw them).
#' @param STAPLEWEX Staple line width expansion, proportional to box width. Default is 0.
#' @param WHISKLWD Whisker line width expansion. Default is 2.
#' @param BOXWEX A scale factor to be applied to all boxes. When there are only a few groups, the
#' appearance of the plot can be improved by making the boxes narrower. Default is 0.7.
#' @param BOXLWD Width of boxplot outer lines. Default is 0.1.
#' @param MEDLWD Width of the median line. Default is 2.6.
#' @param PCH The type of points to be drawn for outliers. Default is 19. See \code{\link[graphics:points]{points}}
#' for possible values and their interpretation.
#' @param ... Any other graphical parameter of `boxplot()`.
#'
#' @return A panel of conditional boxplots.
#' @export
#' @import graphics
#' @importFrom fields bplot.xy
#'
#' @examples
#' cond.boxplots(dataset = iris, Y = 1, Xs = 2:5, outlier = TRUE)
cond.boxplots <- function(dataset, Y, Xs, outlier, N = 6,
MAR = c(2.2,2.1,0.5,1.7), CEX.LAB = 0.9, FONT.LAB = 2, BTY = "n", FG = "gray35",
COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.8, TCL = -0.3,
MGP = c(1.2, 0.4, 0.2), OMA = c(1, 0, 0, 0),
TYPE = "n", BORDER = "moccasin", COL = "gray50", LTY = 1, STAPLEWEX = 0,
WHISKLWD = 2, BOXWEX = 0.5, BOXLWD = 0.1, MEDLWD = 2.6, PCH = 19, ...){
mydata <- as.data.frame(dataset)
if(!is.numeric(mydata[,Y])){
stop("The Y variable must be numeric! Please check the class of your columns before proceeding.")
}
if(!any(sapply(mydata, is.numeric)) && !any(sapply(mydata, is.factor))){
stop("Input dataset should only contain numeric or factor variables! Please check the class of your columns before proceeding.")
}
if(missing(Y)){
stop("Please specify the Y and X(s) variables to be used and respect the required format. See ?cond.boxplots for details.")
}
nam <- names(mydata)
ncol.data <- ncol(mydata[,Xs])
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
minX <- min(Xs)
maxX <- max(Xs)
yyy <- as.matrix(mydata[,Y])
graphics::par(mfrow= c(ncol.adjust,4), mar=MAR, cex.lab=CEX.LAB, font.lab=FONT.LAB, bty=BTY, fg=FG,
col.axis=COL.AXIS, col.lab=COL.LAB, cex=CEX.PAR, tcl=TCL, mgp=MGP, oma=OMA)
for (i in c(minX:maxX)) {
if (is.numeric(mydata[,i])) {
fields::bplot.xy(y = mydata[,Y], x = mydata[,i], outlier=outlier, N=N,
xlab=(nam[i]), type=TYPE, border=BORDER, col=COL,lty=LTY, staplewex=STAPLEWEX,
whisklwd=WHISKLWD, boxwex=BOXWEX, boxlwd=BOXLWD, medlwd=MEDLWD, pch=PCH, cex=0.7, ...)
}else if (is.factor(mydata[,i])) {
xxx <- as.matrix(mydata[,i])
graphics::boxplot(yyy~xxx, outline=outlier, ylab="",
xlab=(nam[i]), type=TYPE, border=BORDER, col=COL, lty=LTY, staplewex=STAPLEWEX,
whisklwd=WHISKLWD, boxwex=BOXWEX, boxlwd=BOXLWD,medlwd=MEDLWD, pch=PCH, cex=0.7, ...)
}
}
}
mrelnoob/jk.dusz.tarping documentation built on July 31, 2023, 9:19 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.
Defines functions cond.boxplots uni.histograms uni.simudistrib uni.dotplots uni.boxplots model_datasets
Documented in cond.boxplots model_datasets uni.boxplots uni.dotplots uni.histograms uni.simudistrib
############################
##### DATA PREPARATION #####
############################
### IMPORTANT NOTE: to avoid creating notes for unquoted variables, I must add the following code at
# the beginning of every source file (e.g. R/myscript.R) that uses unquoted variables, so in front of
# all my scripts doing any kind of analyses (otherwise, I should always assign each variable, e.g.
# mydata$myvariable, which is quite time consuming and wearisome).
if(getRversion() >= "2.15.1") utils::globalVariables(c(
"manager_id", "xp_id", "country", "latitude", "longitude", "elevation", "tarping_date",
"planned_duration", "goals",
"restoration", "operation_type", "multiple_ops", "freq_monitoring", "slope", "difficulty_access",
"shade", "forest", "ruggedness", "granulometry", "obstacles", "flood", "environment", "fabric_type",
"liner_geomem", "agri_geomem", "woven_geotex", "mulching_geotex", "pla_geotex", "weedsp_geotex",
"other_unknown", "grammage", "thickness", "resi_punc", "resi_trac", "season", "maxveg", "preparation",
"levelling", "stand_surface", "age", "fully_tarped", "distance", "multi_strips", "strips_overlap",
"stripsoverlap_ok", "tarpfix_pierced", "geomem", "geotex", "uprootexcav", "efficiency", "high_eff",
"strips_fixation", "staples_distance", "fabric_fixation", "tarpfix_multimethod", "sedicover_height",
"trench", "trench_depth", "pierced_tarpinstall", "plantation", "repairs", "add_control",
"add_control_type", "degradation", "pb_fixation", "pb_durability", "pb_trampiercing", "pb_vandalism",
"regrowth_during", "reg_staples", "reg_stripsoverlap", "reg_obstacles", "reg_holes", "reg_plantations",
"reg_pierced", "reg_edges", "reg_nearby", "untarped_regrowth", "tarping_abandoned", "tarping_completed",
"tarping_ongoing", "tarping_duration", "latest_condition", "latest_regrowth", "latest_months",
"eff_expansion", "eff_dispersal", "eff_vigour", "eff_eradication"))
# Alternatively, here's another solution (to insert within a given function) when the number of unquoted
# variables is low:
# planned_duration <- age <- plantation <- NULL
# Avoids potentially problematic NOTE during the R CMD check (i.e. devtools::check()) due to the
# fact that these 3 variables are un-quoted (also known as non-standard evaluation (NSE)). If you want
# to find again how I know that, just run a Google search with some of the NOTE message "package: no
# visible binding for global variable" (I read a r-blogger.com post about it).
# _________________________________________
### Creation of a function that will create separate datasets for each response variable (depending on the
# explanatory variables their models will potentially include):
#' Datasets for Models Building
#'
#' @description The `model_datasets` function creates the respective reduced datasets that should be used
#' to model each response variables. For instance, if `response.var = "overlaps"`, the function will
#' produce a dataset containing \emph{reg_stripsoverlap} as the \strong{response} variable and a subset of
#' variables to be used as \strong{predictors/explanatory variables} (removing all the variables that
#' should not be used to model \emph{reg_stripsoverlap}'s variations).
#' @param response.var A character string specifying which modelling dataset should be produced (either:
#' efficiency", "edges", "overlaps", or "tarpedarea"):
#' * "efficiency" will produce the dataset for the 3 efficiency evaluation variables (namely \emph{
#' eff_eradication}, \emph{eff_expansion} and \emph{eff_vigour});
#' * "edges" will produce the dataset having for response variable the tarping operations that observed
#' regrowth at the edge of the tarped area (at the end of the operation or during the latest visit to the
#' site);
#' * "overlaps" produce the dataset having for response variable the tarping operations that observed
#' regrowth at strip overlaps;
#' * "tarpedarea" will produce the dataset having for response variable the tarping operations that observed
#' regrowth at somewhere on the area covered by the fabric.
#'
#' @return A tibble.
#' @export
#' @import dplyr stringr
#'
#' @examples
#' \dontrun{
#' eff_model <- model_datasets(response.var = "efficiency")
#' }
model_datasets <- function(response.var = c("efficiency", "edges", "overlaps",
"tarpedarea")){
ppp <- jk.dusz.tarping::clean_my_data()
ppp %>%
dplyr::filter(!stringr::str_detect(operation_type, "crushing_tarping_trial")) %>%
dplyr::mutate(uprootexcav = ifelse(preparation == "excavation" |
preparation == "uprooting", 1, 0)) -> qqq # Exclude rows which
# belong to crushing-tarping trials, and creates a new variable called "uprootexcav"!
### For the 3 "efficiency" models
if (response.var == "efficiency") {
qqq <- dplyr::mutate(.data = qqq,
high_eff = ifelse(stringr::str_detect(latest_regrowth, stringr::regex("few.", dotall = TRUE)) |
eff_eradication == "1", 1, 0)) %>%
dplyr::mutate_if(is_binary, factor)
# With str_detect and regex, I matched all obs. with latest_regrowth that begins with "few"
# (quite similar to grep)!
qqq <- dplyr::mutate(.data = qqq,
efficiency = rowMeans(x = qqq[,80:82], na.rm = FALSE))
tapioca <- qqq[,c("manager_id", "xp_id", "efficiency", "eff_eradication", "high_eff",
"latitude", "longitude", "elevation", "goals",
"freq_monitoring", "slope", "difficulty_access", "shade", "forest", "ruggedness",
"granulometry", "obstacles", "flood",
"geomem", "geotex", "liner_geomem", "agri_geomem", "woven_geotex", "mulching_geotex",
"pla_geotex", "weedsp_geotex", "other_unknown", "grammage", "thickness",
"maxveg", "uprootexcav", "stand_surface", "age", "fully_tarped", "distance", "tarping_duration",
"stripsoverlap_ok", "tarpfix_multimethod", "tarpfix_pierced", "sedicover_height", "trench_depth",
"pierced_tarpinstall", "plantation", "repairs", "add_control", "add_control_type",
"degradation", "pb_fixation", "pb_durability")]
}
### For the "lreg_edges" model
if (response.var == "edges") {
qqq <- dplyr::mutate(.data = qqq, lreg_edges =
ifelse(c(grepl("*edges*", latest_regrowth) | grepl("*edges*", untarped_regrowth)), 1, 0)) %>%
dplyr::mutate(.data = qqq, reg_elsewhere =
ifelse(reg_staples == 1 | reg_stripsoverlap == 1 | reg_obstacles == 1 |
reg_holes == 1 | reg_plantations == 1 | reg_pierced == 1, yes = 1, no = 0)) %>%
dplyr::filter(fully_tarped == 1) # Only for fully tarped operations!
tapioca <- qqq[,c("manager_id", "xp_id", "lreg_edges",
"latitude", "longitude", "elevation", "slope", "flood", "difficulty_access", "shade", "forest",
"ruggedness", "granulometry", "obstacles", "stand_surface",
"geomem", "geotex",
"maxveg", "uprootexcav",
"distance", "stripsoverlap_ok",
"tarpfix_multimethod", "tarpfix_pierced", "sedicover_height", "trench_depth", "plantation",
"repairs", "add_control", "freq_monitoring",
"degradation", "pb_fixation", "pb_durability", "tarping_duration", "reg_elsewhere")]
}
### For the "overlaps" model
if (response.var == "overlaps") {
qqq <- dplyr::mutate(.data = qqq, reg_elsewhere =
ifelse(reg_staples == 1 | reg_edges == 1 | reg_obstacles == 1 |
reg_holes == 1 | reg_plantations == 1 | reg_pierced == 1, yes = 1, no = 0)) %>%
dplyr::mutate(stripfix_pierced =
ifelse(c(grepl("*staples*", strips_fixation) | grepl("*stakes*", strips_fixation)), 1, 0)) %>%
dplyr::mutate(stripfix_taped =
ifelse(c(strips_fixation == "tape_and_heavyobj" | strips_fixation == "double_tape" |
strips_fixation == "tape" | strips_fixation == "tape_and_staples"), 1, 0)) %>%
dplyr::filter(multi_strips == 1) %>%
dplyr::mutate_if(is_binary, factor)
tapioca <- qqq[,c("manager_id", "xp_id", "reg_stripsoverlap",
"latitude", "longitude", "elevation", "slope", "flood", "difficulty_access","shade", "forest", "ruggedness",
"granulometry", "obstacles", "stand_surface",
"geomem", "geotex",
"maxveg", "uprootexcav", "levelling", "fully_tarped", "distance",
"strips_overlap", "tarpfix_multimethod", "tarpfix_pierced", "stripfix_pierced", "stripfix_taped",
"sedicover_height", "trench_depth", "plantation",
"repairs", "add_control", "freq_monitoring", "pb_fixation", "pb_durability",
"reg_elsewhere", "tarping_duration")]
}
### For the "tarpedarea" model
if (response.var == "tarpedarea") {
qqq <- dplyr::mutate(.data = qqq, lreg_tarpedarea =
ifelse(c(grepl("*tarpedarea*", latest_regrowth) |
grepl("*tarpedarea*", untarped_regrowth)), 1, 0))
tapioca <- qqq[,c("manager_id", "xp_id", "lreg_tarpedarea",
"latitude", "longitude", "elevation", "slope", "flood",
"difficulty_access", "shade", "forest", "ruggedness", "granulometry", "obstacles", "stand_surface",
"geomem", "geotex", "woven_geotex",
"maxveg", "uprootexcav", "levelling", "fully_tarped", "distance",
"stripsoverlap_ok", "tarpfix_multimethod", "tarpfix_pierced", "pierced_tarpinstall",
"sedicover_height", "trench_depth", "plantation",
"repairs", "add_control", "freq_monitoring",
"pb_fixation", "pb_durability", "pb_trampiercing", "reg_edges", "tarping_duration")]
}
return(tapioca)
}
# _________________________________________
### Creation of a function that draws univariate boxplots for all numeric variables in a given dataset:
#' Univariate boxplots
#'
#' @description The `uni.boxplots` function draws, within a single panel, an independent boxplot for each
#' numeric (continuous or discrete) variable in a given dataset. It is particularly useful for data
#' exploration (e.g. Zuur \emph{et al.}, 2010). For instance, to simultaneously observe the
#' distributions of all numeric variables or \strong{to detect their univariate outliers}.
#'
#' @details The `uni.boxplots` function only modifies the graphical parameters of the
#' \code{\link[graphics:boxplot]{boxplot}} function in the `graphics` package to match some predefined
#' preferences. Therefore, default values of `uni.boxplots` create nice looking boxplots but retain
#' default \emph{heuristic} aspects of `boxplot` (such as the length of whiskers or the plotting of
#' outliers). These aspects can however be changed as in \code{\link[graphics:boxplot]{boxplot}}. \cr
#' On the other hand, panel parameters are internally controlled using `par`. However, to avoid unforeseen
#' conflicts with other internal parameters, it is not possible to tune panel parameters as we would
#' do with `par`. Instead, parametrization is only possible with the given subset of parameters.
#'
#' @note To avoid \emph{recursive argument errors}, internal arguments should be called using upper case
#' letters (e.g. CEX.LAB = 0.9) whereas other arguments from the `boxplot` function should be called with
#' their normal case writing (e.g. outline = FALSE)!
#'
#' @param dataset The input dataset containing all variables to be plotted. It may contain all kinds of
#' variables, the `uni.boxplots` function will automatically detect and plot numeric variables (columns).
#' @param ... Any other parameter that can be incorporated in \code{\link[graphics:boxplot]{boxplot}}.
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(0.5,4.1,1.1,1.5)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting of `cex`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(2.4, 0.6, 0).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param TYPE The type of boxplot to draw. Default is "n".
#' @param BORDER An optional vector of colors for the outlines of the boxplots. The values in border
#' are recycled if the length of border is less than the number of plots. Default is "lightcoral".
#' @param COL If col is non-null it is assumed to contain colors to be used to colour the bodies of
#' the boxplots. Default is "moccasin".
#' @param LTY The line type. Line types can either be specified as an integer (0=blank, 1=solid (default),
#' 2=dashed, 3=dotted, 4=dotdash, 5=longdash, 6=twodash) or as one of the character strings "blank",
#' "solid", "dashed", "dotted", "dotdash", "longdash", or "twodash", where "blank" uses ‘invisible lines’
#' (i.e., does not draw them).
#' @param STAPLEWEX Staple line width expansion, proportional to box width. Default is 0.
#' @param WHISKLWD Whisker line width expansion. Default is 2.
#' @param BOXWEX A scale factor to be applied to all boxes. When there are only a few groups, the
#' appearance of the plot can be improved by making the boxes narrower. Default is 0.7.
#' @param BOXLWD Width of boxplot outer lines. Default is 0.1.
#' @param MEDLWD Width of the median line. Default is 2.6.
#' @param PCH The type of points to be drawn for outliers. Default is 19. See \code{\link[graphics:points]{points}}
#' for possible values and their interpretation.
#'
#' @return A panel of univariate boxplots.
#' @export
#' @import graphics
#'
#' @examples
#' data("mtcars")
#' uni.boxplots(dataset = mtcars)
uni.boxplots <- function(dataset, MAR=c(0.5,4.1,1.1,1.5), CEX.LAB=1, FONT.LAB=2, BTY = "n", FG = "gray35",
COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.8, TCL = -0.3,
MGP = c(2.4, 0.6, 0), OMA = c(1, 0, 0, 0),
TYPE = "n", BORDER = "lightcoral", COL = "moccasin", LTY = 1, STAPLEWEX = 0,
WHISKLWD = 2, BOXWEX = 0.7, BOXLWD = 0.1, MEDLWD = 2.6, PCH = 19, ...){
num.data <- dataset[, sapply(dataset, is.numeric)]
nam <- names(num.data)
ncol.data <- ncol(num.data)
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
graphics::par(mfrow= c(ncol.adjust,4), mar=MAR, cex.lab=CEX.LAB, font.lab=FONT.LAB, bty=BTY, fg=FG,
col.axis=COL.AXIS, col.lab=COL.LAB, cex=CEX.PAR, tcl=TCL, mgp=MGP, oma=OMA)
for (i in c(1:ncol.data)) {
graphics::boxplot(num.data[,i],ylab =(nam[i]), type=TYPE, border=BORDER, col = COL,
lty=LTY, staplewex=STAPLEWEX, whisklwd=WHISKLWD, boxwex=BOXWEX, boxlwd=BOXLWD,
medlwd=MEDLWD, pch=PCH, cex=0.7, ...) }
}
# _________________________________________
### Creation of a function that draws univariate Cleveland dotplots for all numeric variables in a
### given dataset:
#' Univariate Cleveland dotplots
#'
#' @description The `uni.dotplots` function draws, within a single panel, an independent Cleveland dotplot
#' (i.e. plotting value against rank) for each numeric (continuous or discrete) variable in a given
#' dataset. It is particularly useful for data exploration (e.g. Zuur \emph{et al.}, 2010). For instance,
#' to simultaneously observe the distributions of all numeric variables or \strong{to detect their
#' univariate outliers}.
#'
#' @details The `uni.dotplots` function only modifies the graphical parameters of the
#' \code{\link[graphics:plot.default]{plot}} function in the `graphics` package to match some predefined
#' preferences. Therefore, default values of `uni.dotplots` create nice looking dotplots but retain
#' some aspects of the original `plot` function. These aspects can however be changed as in
#' \code{\link[graphics:plot.default]{plot}}. \cr
#' On the other hand, panel parameters are internally controlled using `par`. However, to avoid unforeseen
#' conflicts with other internal parameters, it is not possible to tune panel parameters as we would
#' do with `par`. Instead, parametrization is only possible with the given subset of parameters.
#'
#' @note To avoid \emph{recursive argument errors}, internal arguments should be called using upper case
#' letters (e.g. CEX.LAB = 0.9) whereas other arguments from the `plot` function should be called with
#' their normal case writing (e.g. sub = "My subtitle")!
#'
#' @param dataset The input dataset containing all variables to be plotted (must be a `data.frame` with
#' at least 2 variables). It may contain all kinds of columns, the `uni.dotplots` function will
#' automatically detect and plot numeric variables (columns).
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(0.5,4.1,1.1,1.5)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting
#' of `CEX.PAR`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(2.4, 0.6, 0).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param LAB A numerical vector of the form `c(x, y, len)` which modifies the default way that axes
#' are annotated. The values of `x` and `y` give the (approximate) number of tickmarks on the x and y
#' axes and len specifies the label length. The default is `c(5, 5, 7)`. Note that this only affects
#' the way the parameters xaxp and yaxp are set when the user coordinate system is set up, and is
#' not consulted when axes are drawn. `len` is unimplemented in `R`.
#' @param COL.PCH The color to be used for points. Default is "lightcoral".
#' @param PCH The type of points to be drawn. Default is 19. See \code{\link[graphics:points]{points}}
#' for possible values and their interpretation.
#' @param COL.GRID The color of the background grid. Default is "lavender".
#' @param NX The number of lines for the grid in x. Default value is 5.
#' @param NY The number of lines for the grid in Y. Default value is 9.
#' @param LTY The type of lines to be drawn in the background grid. Default value is 6.
#' @param ... Any other parameter that can be incorporated in \code{\link[graphics:plot.default]{plot}}.
#'
#' @return A panel of univariate dotplots.
#' @export
#' @import graphics
#'
#' @examples
#' data("mtcars")
#' uni.dotplots(dataset = mtcars, COL.GRID = "lightblue", LTY = 1, NX = 10, NY = 20)
uni.dotplots <- function(dataset, MAR=c(3,2,0.5,1.5), CEX.LAB = 1.2, FONT.LAB = 2, BTY = "n",
FG = "gray35", COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.6,
TCL = -0.3, MGP = c(1.7, 0.6, 0.1), OMA = c(1, 0, 1, 0), LAB = c(5, 10, 7),
COL.PCH = "lightcoral", PCH = 19, COL.GRID = "lavender", NX = 5, NY = 9, LTY = 6,
...){
num.data <- dataset[, sapply(dataset, is.numeric)]
nam <- names(num.data)
ncol.data <- ncol(num.data)
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
num.data <- as.matrix(num.data)
graphics::par(mfrow= c (ncol.adjust,4), mar=MAR, cex.lab = CEX.LAB, font.lab=FONT.LAB, bty = BTY, fg = FG,
col.axis = COL.AXIS, col.lab = COL.LAB, cex = CEX.PAR, tcl = TCL,
mgp = MGP, oma = OMA, lab = LAB)
for (i in c(1:ncol(num.data))) {
graphics::plot(x = num.data[,i], y = 1:length(num.data[,i]), type = "p", xlab = nam[i], ylab = "",
col = COL.PCH, pch = PCH, panel.first = {
grid(col=COL.GRID,nx = NX,ny = NY, lty = LTY)
}, ...) }
# Here, the argument panel.first={} is used to draw the grid first, so behind the points!
}
# _________________________________________
### Creation of a function that generate random samples from various distributions based on my variables
### parameters:
#' Random sample simulation
#'
#' @description The `uni.simudistrib` function **automatically generates 5 Cleveland dotplots** of random
#' samples from different distributions (either `normal`, `log-normal`, or `poisson`) based on the
#' parameters of the variables in the input `data.frame` or `matrix` (see \emph{Details}). \cr
#' This function is useful to see whether variables' extreme values are actual outliers or whether they
#' lie in a range of values possible for a random sample drawn from a `normal`, `log-normal`, or a `poisson`
#' distribution. \emph{In fine}, it may help determine if the original variable can be approximated by
#' these distribution with or without a transformation.
#'
#' @details The `uni.simudistrib` function extracts some key parameters from the input variables (sample
#' size, mean and standard deviation) and generates random samples based on these parameters. For instance,
#' if `simu.var` contains \emph{i} variables `X1`, `X2`, ... `Xi` and if `distribution = "normal"`, the
#' function will return a panel of \emph{i}x5 plots:
#'
#' * The 1st row will contain five dotplots for five random samples with \emph{n} = `length(X1)` and drawn
#' from a Normal distribution with the same mean and standard deviation as `X1`.
#' * The 2nd row will contain five dotplots for five random samples with \emph{n} = `length(X2)` and drawn
#' from a Normal distribution with the same mean and standard deviation as `X2`.
#' * Etc.
#'
#' \strong{Warning}: the function may fail for `log-normal` and `poisson` distributions if input
#' variables contain negative values (because these distributions are by definition positive). Additionally,
#' if `distribution = "poisson"`, the resulting plots will return \strong{integer} values as Poisson
#' is a discrete probability distribution.
#'
#' @param simu.var A `data.frame` or a `matrix`. For obvious layout and readability reasons, `simu.var`
#' should not include too many variables (`p` < 12 is advised) if the plot is to be printed in a page
#' or window of limited dimensions. For any use in HTML documents (e.g. with `RMarkdown`), the number of
#' input variables should not be a problem.
#' @param distribution Either `"normal"`, `"log-normal"` or `"poisson"` (case sensitive).
#'
#' @return A panel of `p`*5 plots, where `p` is the number of variables in `simu.var`.
#' @export
#' @import graphics
#' @importFrom stats na.omit sd rnorm rlnorm rpois
#'
#' @examples
#' uni.simudistrib(simu.var = iris[,1:4], distribution = "normal")
uni.simudistrib <- function(simu.var, distribution){
if (!is.data.frame(simu.var) && !is.matrix(simu.var)) {
warning("The input dataset (i.e. simu.var) must either be a data.frame or a matrix!")
}
if(missing(distribution)){
stop("Please specify the desired distribution. See ?uni.simudistrib for details.")
}
ncol.data <- ncol(simu.var)
nam <- names(simu.var)
num.data <- as.matrix(simu.var)
graphics::par(mfrow = c(ncol.data,5), mar = c(3,2,0.1,1.5), cex.lab = 1, font.lab=2, bty = "n",
fg = "gray35", col.axis = "gray35", col.lab = "gray20", cex = 0.6, tcl = -0.3,
mgp = c(1.8, 0.6, 0.1), oma = c(0.2, 0.2, 0, 0), lab = c(5, 10, 7))
for (i in c(1:ncol(num.data))) {
mu <- mean(stats::na.omit(num.data[,i]))
std <- stats::sd(stats::na.omit(num.data[,i]))
n <- length(num.data[,i])
rrr <- NULL
if (distribution == "normal") {
for (j in 1:5) {
rrr[[j]] <- stats::rnorm(n = n, mean = mu, sd = std)
graphics::plot(x = rrr[[j]], y = 1:length(rrr[[j]]), type = "p",
xlab = paste("rnorm based on", nam[i], sep = " "), ylab = "", col = "hotpink3",
pch = 19, panel.first = {grid(col="lavender",nx = 9,ny = 3, lty = 6)})}
}
if (distribution == "log-normal") {
for (j in 1:5) {
rrr[[j]] <- stats::rlnorm(n = n,
meanlog = log(mu^2 / sqrt(std^2 + mu^2)),
sdlog = sqrt(log(1 + (std^2 / mu^2)))) # That's how you use rlnorm!!!!!!
graphics::plot(x = rrr[[j]], y = 1:length(rrr[[j]]), type = "p",
xlab = paste("rlnorm based on", nam[i], sep = " "), ylab = "", col = "hotpink3",
pch = 19, panel.first = {grid(col="lavender",nx = 9,ny = 3, lty = 6)})}
}
if (distribution == "poisson") {
for (j in 1:5) {
rrr[[j]] <- stats::rpois(n = n, lambda = mu)
graphics::plot(x = rrr[[j]], y = 1:length(rrr[[j]]), type = "p",
xlab = paste("rpois based on", nam[i], sep = " "), ylab = "", col = "hotpink3",
pch = 19, panel.first = {grid(col="lavender",nx = 9,ny = 3, lty = 6)})}
}
}
}
# _________________________________________
### Creation of a function that generates histograms for each numeric variable:
#' Histogram Panel
#'
#' @description The `uni.histograms` function draws, within a single panel, an independent histogram
#' or each numeric (continuous or discrete) variable in a given dataset. It is particularly useful
#' for data exploration (e.g. Zuur \emph{et al.}, 2010). For instance, to simultaneously observe
#' the distributions of all numeric variables and determine which one will require transformation.
#'
#' @param dataset The input dataset containing all variables to be plotted (must be a `data.frame` with
#' at least 2 variables). It may contain all kinds of columns, the `uni.histograms` function will
#' automatically detect and plot numeric variables (columns).
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(0.5,4.1,1.1,1.5)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting
#' of `CEX.PAR`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(2.4, 0.6, 0).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param LAB A numerical vector of the form `c(x, y, len)` which modifies the default way that axes
#' are annotated. The values of `x` and `y` give the (approximate) number of tickmarks on the x and y
#' axes and len specifies the label length. The default is `c(5, 5, 7)`. Note that this only affects
#' the way the parameters xaxp and yaxp are set when the user coordinate system is set up, and is
#' not consulted when axes are drawn. `len` is unimplemented in `R`.
#' @param BREAKS One of:
#' * a vector giving the breakpoints between histogram cells,
#' * a function to compute the vector of breakpoints,
#' * a single number giving the number of cells for the histogram,
#' * a character string naming an algorithm to compute the number of cells (see
#' \code{\link[graphics:hist]{hist}}),
#' * a function to compute the number of cells.
#'
#' In the last three cases the number is a suggestion only; as the breakpoints will be set to pretty
#' values, the number is limited to 1e6 (with a warning if it was larger). If breaks is a function,
#' the x vector is supplied to it as the only argument (and the number of breaks is only limited by
#' the amount of available memory).
#' @param COL The color of the bar of the histograms (bins).
#' @param BORDER The color of the border of the bars.
#'
#' @return A panel of histograms.
#' @import graphics
#' @export
#'
#' @examples
#' uni.histograms(dataset = iris[,1:4])
uni.histograms <- function(dataset, MAR=c(3,2,0.5,1.5), CEX.LAB = 1.2, FONT.LAB = 2, BTY = "n",
FG = "gray35", COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.6,
TCL = -0.3, MGP = c(1.7, 0.6, 0.1), OMA = c(1, 0, 1, 0), LAB = c(5, 10, 7),
BREAKS = 10, COL = "moccasin", BORDER = "white"){
num.data <- dataset[, sapply(dataset, is.numeric)]
nam <- names(num.data)
ncol.data <- ncol(num.data)
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
num.data <- as.matrix(num.data)
graphics::par(mfrow= c (ncol.adjust,4), mar=MAR, cex.lab = CEX.LAB, font.lab=FONT.LAB, bty = BTY, fg = FG,
col.axis = COL.AXIS, col.lab = COL.LAB, cex = CEX.PAR, tcl = TCL,
mgp = MGP, oma = OMA, lab = LAB)
for (i in c(1:ncol(num.data))) {
graphics::hist(num.data[,i], breaks = BREAKS, col = COL, border = BORDER,
main = "", xlab = nam[i], ylab = "")
}
}
# ___________________________________________________________
### Creation of a function that generates conditional boxplots between the given variables:
#' Bivariate Conditional Boxplots
#'
#' @description The `cond.boxplots` function draws, within a single panel, conditional boxplots (i.e.
#' a plot of a continuous Y variable against one or several categorical X(s) variable(s) (factor),
#' or \strong{discretized} numeric variable(s)) in order to explore bivariate relationships and
#' assess \strong{heteroscedasticity}. \cr
#' If some of the X variables are quantitative (numeric), they will consequently be transformed into
#' factors and discretized into a given number of classes, set approximately by the `N` parameter. This
#' factorisation is required because boxplots are not meant to plot two quantitative variables (without
#' factorisation, the function would plot as many boxplots as there are values in X).
#'
#' @param dataset The input `data.frame` containing the variables to plot (both the Y and the X(s)). Must
#' only contain numeric or factor variables!
#' @param Y The number of the column of `dataset` containing the variable to be plotted as Y. This
#' parameter should be specified as an integer. For instance, if the variable to be used as Y is called
#' "blip" and is the second column of the `dataset`, the `Y` parameter should be `2` (and NOT "two",
#' `dataset$blip`, or `dataset[,2]`).
#' @param Xs The numbers of the columns to be plotted as Xs (e.g. `4:10`, if all the columns from the
#' fourth to the tenth are to be plotted against `Y`).
#' @param outlier Logical (`TRUE` or `FALSE`). Should outliers be plotted or not?
#' @param N Integer. The number of bins on X. Default is 6.
#'
#' @param MAR A numerical vector of the form `c(bottom, left, top, right)` which gives the number of lines
#' of margin to be specified on the four sides of the plot. The default is `c(2.2,2.1,0.5,1.7)`.
#' @param CEX.LAB The magnification to be used for x and y labels relative to the current setting
#' of `CEX.PAR`.
#' @param FONT.LAB The font to be used for x and y labels.
#' @param BTY A character string which determined the type of box which is drawn about plots. If `BTY` is
#' one of "o", "l", "7", "c", "u", or "]" the resulting box resembles the corresponding upper
#' case letter. A value of "n" suppresses the box (the default).
#' @param FG The color to be used for the foreground of plots. This is the default color used for things
#' like axes and boxes around plots (defaults to "gray35").
#' @param COL.AXIS The color to be used for axis annotation. Defaults to "gray35".
#' @param COL.LAB The color to be used for x and y labels. Defaults to "gray20".
#' @param CEX.PAR A numerical value giving the amount by which plotting text and symbols should be
#' magnified relative to the default (for `par`, the panel manager). This starts as 1 when a device
#' is opened, and is reset when the layout is changed, e.g. by setting `mfrow`. Defaults to 0.8.
#' @param TCL The length of tick marks as a fraction of the height of a line of text. The default
#' value is -0.3.
#' @param MGP The margin line (in `mex` units) for the axis title, axis labels and axis line.
#' Note that `mgp[1]` affects title whereas `mgp[2:3]` affect axis. The default is c(1.2, 0.4, 0.2).
#' @param OMA A vector of the form `c(bottom, left, top, right)` giving the size of the outer margins
#' in lines of text.
#' @param TYPE The type of boxplot to draw. Default is "n".
#' @param BORDER An optional vector of colors for the outlines of the boxplots. The values in border
#' are recycled if the length of border is less than the number of plots. Default is "lightcoral".
#' @param COL If col is non-null it is assumed to contain colors to be used to colour the bodies of
#' the boxplots. Default is "moccasin".
#' @param LTY The line type. Line types can either be specified as an integer (0=blank, 1=solid (default),
#' 2=dashed, 3=dotted, 4=dotdash, 5=longdash, 6=twodash) or as one of the character strings "blank",
#' "solid", "dashed", "dotted", "dotdash", "longdash", or "twodash", where "blank" uses ‘invisible lines’
#' (i.e., does not draw them).
#' @param STAPLEWEX Staple line width expansion, proportional to box width. Default is 0.
#' @param WHISKLWD Whisker line width expansion. Default is 2.
#' @param BOXWEX A scale factor to be applied to all boxes. When there are only a few groups, the
#' appearance of the plot can be improved by making the boxes narrower. Default is 0.7.
#' @param BOXLWD Width of boxplot outer lines. Default is 0.1.
#' @param MEDLWD Width of the median line. Default is 2.6.
#' @param PCH The type of points to be drawn for outliers. Default is 19. See \code{\link[graphics:points]{points}}
#' for possible values and their interpretation.
#' @param ... Any other graphical parameter of `boxplot()`.
#'
#' @return A panel of conditional boxplots.
#' @export
#' @import graphics
#' @importFrom fields bplot.xy
#'
#' @examples
#' cond.boxplots(dataset = iris, Y = 1, Xs = 2:5, outlier = TRUE)
cond.boxplots <- function(dataset, Y, Xs, outlier, N = 6,
MAR = c(2.2,2.1,0.5,1.7), CEX.LAB = 0.9, FONT.LAB = 2, BTY = "n", FG = "gray35",
COL.AXIS = "gray35", COL.LAB = "gray20", CEX.PAR = 0.8, TCL = -0.3,
MGP = c(1.2, 0.4, 0.2), OMA = c(1, 0, 0, 0),
TYPE = "n", BORDER = "moccasin", COL = "gray50", LTY = 1, STAPLEWEX = 0,
WHISKLWD = 2, BOXWEX = 0.5, BOXLWD = 0.1, MEDLWD = 2.6, PCH = 19, ...){
mydata <- as.data.frame(dataset)
if(!is.numeric(mydata[,Y])){
stop("The Y variable must be numeric! Please check the class of your columns before proceeding.")
}
if(!any(sapply(mydata, is.numeric)) && !any(sapply(mydata, is.factor))){
stop("Input dataset should only contain numeric or factor variables! Please check the class of your columns before proceeding.")
}
if(missing(Y)){
stop("Please specify the Y and X(s) variables to be used and respect the required format. See ?cond.boxplots for details.")
}
nam <- names(mydata)
ncol.data <- ncol(mydata[,Xs])
ncol.adjust <- ceiling(x = ncol.data/4) # Round to the next integer (e.g. ceiling(x = 7.12) returns 8)!
minX <- min(Xs)
maxX <- max(Xs)
yyy <- as.matrix(mydata[,Y])
graphics::par(mfrow= c(ncol.adjust,4), mar=MAR, cex.lab=CEX.LAB, font.lab=FONT.LAB, bty=BTY, fg=FG,
col.axis=COL.AXIS, col.lab=COL.LAB, cex=CEX.PAR, tcl=TCL, mgp=MGP, oma=OMA)
for (i in c(minX:maxX)) {
if (is.numeric(mydata[,i])) {
fields::bplot.xy(y = mydata[,Y], x = mydata[,i], outlier=outlier, N=N,
xlab=(nam[i]), type=TYPE, border=BORDER, col=COL,lty=LTY, staplewex=STAPLEWEX,
whisklwd=WHISKLWD, boxwex=BOXWEX, boxlwd=BOXLWD, medlwd=MEDLWD, pch=PCH, cex=0.7, ...)
}else if (is.factor(mydata[,i])) {
xxx <- as.matrix(mydata[,i])
graphics::boxplot(yyy~xxx, outline=outlier, ylab="",
xlab=(nam[i]), type=TYPE, border=BORDER, col=COL, lty=LTY, staplewex=STAPLEWEX,
whisklwd=WHISKLWD, boxwex=BOXWEX, boxlwd=BOXLWD,medlwd=MEDLWD, pch=PCH, cex=0.7, ...)
}
}
}
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.