biplot_2d: 2D biplot
In Andros-Spica/biplot2d3d: Create and Customize Ordination Plots (Biplots) in 2D and 3D
Description
Usage
Arguments
Details
Examples
Description
Generates a 2D biplot representing the default
points/scores and loadings of an ordination object,
such as a PCA produced by princomp.
Usage
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
biplot_2d(ordination_object, ordination_method = "PCA",
biplot_type = "default", rows_over_columns = 0.5, groups = NULL,
vips = NULL, detach_arrows = TRUE, show_grid = TRUE, show_axes = TRUE,
show_group_legend = FALSE, show_vip_legend = TRUE, show_arrows = TRUE,
show_fitAnalysis = TRUE, main_lwd = 2, grid_cex = 1, grid_font = 3,
grid_adj = 0.5, invert_coordinates = c(FALSE, FALSE), xlim = NULL,
ylim = NULL, x_title = "", y_title = "", x_title_cex = 1,
x_title_font = 2, y_title_cex = 1, y_title_font = 2, subtitle = NULL,
subtitle_position = "bottomleft", subtitle_cex = 1.2,
point_type = "point", point_pch = 1, point_size = 1,
point_label = NULL, point_label_cex = 1, point_label_font = 3,
point_label_adj = c(0.5, 0.5), point_label_adj_override = NULL,
arrow_color = "darkorange", arrow_mim_dist = 0, arrow_length = 0.2,
arrow_cex = 0.1, arrow_lwd = 2, arrow_label_cex = 1,
arrow_label_color = "black", arrow_label_font = 1,
arrow_label_adj = c(0.5, 0.5), arrow_label_adj_override = NULL,
group_color = "black", group_star_cex = 1, group_ellipse_cex = 1,
group_ellipse_axes = FALSE, group_label_cex = 1,
group_legend_title = "groups", group_legend_title_pos = c(0.5, 0.85),
group_legend_title_cex = 1, group_legend_title_font = 3,
group_legend_title_adj = 0.5, group_legend_box_color = "white",
group_legend_key_pch = 15, group_legend_key_cex = 1,
group_legend_key_lwd = 1, group_legend_key_margin = 0.15,
group_legend_text_margin = 0.25, group_legend_text_color = "black",
group_legend_text_cex = 1, group_legend_text_font = 1,
group_legend_text_adj = 0, vip_color = "black", vip_pch = c(0, 1, 5, 2,
6, 4, 3), vip_cex = c(5, 5, 5, 5, 5, 3, 3), vip_lwd = 5, vip_font = 1,
vip_adj = c(0.5, 0.5), vip_legend_title = "VIPs",
vip_legend_title_pos = c(0.5, 0.85), vip_legend_title_cex = 1,
vip_legend_title_font = 3, vip_legend_title_adj = 0.5,
vip_legend_box_color = "white", vip_legend_key_cex = 0.8,
vip_legend_key_margin = 0.15, vip_legend_text_cex = 1,
vip_legend_text_font = 1, vip_legend_text_adj = 0,
vip_legend_text_margin = 0.25, fitAnalysis_lwd = 3,
fitAnalysis_screePlot_color = c("grey", "white"),
fitAnalysis_stress_cex = 1, fitAnalysis_stress_lab_cex = 1,
fitAnalysis_stress_axis_cex = 1, fitAnalysis_stress_p_color = "darkgrey",
fitAnalysis_stress_l_color = "black", test_text = NULL,
test_spacing_paragraph = 0.8, test_spacing_line = 0.8, test_cex = 1,
test_font = 1, test_adj = 0.5, fit_into_main = FALSE, main_fig = c(0,
1, 0, 1), group_legend_fig = c(0.8, 0.99, 0.6, 0.9),
vip_legend_fig = c(0.78, 0.99, 0.3, 0.55), arrow_fig = c(0.69, 0.99, 0.01,
0.31), fitAnalysis_fig = c(0.02, 0.35, 0.06, 0.25), test_fig = c(0, 0.3,
0.8, 0.99), x_title_fig = c(0.25, 1, 0.85, 1), y_title_fig = c(0.91, 1, 0,
1), output_type = c("preview", "png"), open_new_device = TRUE,
leave_device_open = FALSE, directory = "", file_name = "2D Biplot",
width = 400, height = 400, family = "sans")
Arguments
ordination_object
A R object containing a
direct and named reference to default ordination
outputs (i.e. ordination_object$scores or
ordination_object$points,
ordination_object$loadings and
ordination_object$sdev) available for at
least 2 dimensions. Alternatively, a data frame
or matrix with at least two columns is accepted
provided that ordination_method = NULL),
which creates a two-dimensional scatter plot.
ordination_method
Character, the ordination
method that was used to generate the ordination
object: "PCA" for Principal Components Analysis
(default), "PCoA" for Principal Coordinates Analysis,
"NMDS" for Non-metric Multidimensional Scaling, and
"LDA" for Linear Discriminant Analysis.
biplot_type
Character, indicating the type of
biplot transformation of data: "default" and
"pc.biplot", corresponding to the transformations
performed in biplot.princomp
with pc.biplot = FALSE ("default") and
pc.biplot = TRUE ("pc.biplot"). If NULL,
no processing is performed, assuming that data
within ordination_object was previously transformed.
rows_over_columns
Numeric, the value defining the
degree in which distances between observations have
priority over distances between variables
(0 = observation-focused, 1 = variable-focused).
It corresponds to the argument scale in
biplot.princomp. It will be
ignored if biplot_type = NULL. The default is
set at 0.5, which corresponds to a 'SQRT-Biplot',
a compromise between the two representations.
groups
A factor variable containing the group
assignation of each point.
vips
A list of logical (Boolean) vectors identifying
"Very Important Points" under different methods or criteria.
detach_arrows
Logical, wheter to display covariance arrows
a independent miniature plot, overlapping the main plot and
placed according to arrow_fig.
show_grid
Logical, wheter to display the background grid.
show_axes
Logical, wheter to display the axes.
show_group_legend
Logical, whether to display
a legend for groups.
show_vip_legend
Logical, whether to display a
legend for vip criteria.
show_arrows
Logical, whether to show variable
covariance arrows.
show_fitAnalysis
Logical, whether to display the
fit analysis plot corresponding to the ordination method
given (Scree plot or Shepard plot).
main_lwd
The line width to be used in the main plot.
(lwd in (par).
grid_cex, grid_font, grid_adj
The scale, font type,
and text justification of the grid notation.
invert_coordinates
Logical, vector of length two
expressing which dimensions, if any, must be inverted
before plotting (e.g. for aesthetical reasons).
xlim, ylim
the ranges to be encompassed by the x and
y axes, if NULL they are computed
(See s.class).
x_title, y_title
Character, texts to be placed in
the x and y axes.
x_title_cex, x_title_font
The text parameters
of the x axis or horizontal title (par).
y_title_cex, y_title_font
The text parameters
of the y axis or vertical title (par).
subtitle
Character, a subtitle to be displayed
in the bottom-left corner of the plot
(csub in s.class).
If equals NULL and ordination_method = "PCA",
a default subtitle is the R-Squared of the 2D fit
respect to the original distances.
subtitle_position
Character, value indicating the
position of the subtitle in the main plot ("topleft",
"topright", "bottomleft", and "bottomright"; see
s.class).
subtitle_cex
the font size of the
subtitle (csub in
s.class).
point_type
A character input accepting three
values: "point", "label" (displaying the content
of point_label) and "point and label", placing
both points and labels.
point_pch
A number or a numerical vector given
to pch in par.
point_size
The size or scale given to cpoint in
s.class.
point_label
A character vector labelling every
observation. It's length must be equal to the number
of rows in the points/scores of the ordination object
(nrow(ordination_object$points) ==
length(point_label)).
point_label_cex, point_label_font, point_label_adj
The text parameters of the points' labels
(text).
point_label_adj_override
A data frame with x,y
values to be passed to adj, overrinding the
default value given in point_label_adj.
The rows must be named exactly as points
are referred in the ordination object.
arrow_color
The color or colors to be used in
covariance arrows.
arrow_mim_dist
Numeric, the minimum distance of a arrow
from the origin of arrows (i.e. zero covariance),
in order for it to be displayed (range [0 = all arrows
are displayed,1 = no arrow is displayed]).
arrow_length
Numeric, the scalar factor applied to loadings
to resize them respect to scores. If detach_arrows = TRUE,
resizing is controlled with arrow_fig, so this value is ignored.
arrow_cex, arrow_lwd, arrow_label_cex, arrow_label_color, arrow_label_font, arrow_label_adj
Graphical parameters of the covariance arrows and their
labels (see arrows and
par). arrow_cex is actually
equivalent to length in arrows.
arrow_label_adj_override
A data frame with x,y values
to be passed to adj, overrinding the default
value given in arrow_label_adj. The rows must
be named exactly as variables are referred in the
ordination object.
group_color
A vector containing the colors to
be used in each group (applied to points, labels,
stars and ellipses). If NULL, automatically assign
different colors from the rainbow() palette.
group_star_cex, group_ellipse_cex, group_label_cex
The
size or scale of the stars, ellipses, and labels
representing groups passed to s.class
of the ade4 package. Zero values render these
elements invisible.
group_ellipse_axes
Logical, wheter to show
the ellipses axes, passed to axesell in
s.class of the ade4 package.
group_legend_title
Character, the title of the
group legend. If equal NULL or "", no title is displayed.
group_legend_title_pos
A numeric vector of length
two, the xy position of the title within the group legend
box.
group_legend_title_cex, group_legend_title_font, group_legend_title_adj
The text parameters to be applied in the group legend
title (par).
group_legend_box_color
The background color of
the group legend box.
group_legend_key_pch, group_legend_key_cex, group_legend_key_lwd
The type, size and line width of the keys in the
group legend.
group_legend_key_margin
The x position of the keys
within the group legend box. Values from 0 to 1.
group_legend_text_margin
The x position of the
text entries in the group legend. Values from 0 to 1.
group_legend_text_color
The color or colors of
the text entries in the group legend.
group_legend_text_cex, group_legend_text_font, group_legend_text_adj
The text parameters of the text entries in the group legend.
vip_color
The color or colors to be used in vip markings.
vip_pch
A character vector containing the
characters used for the vips markings under
each criterion.
vip_cex, vip_lwd, vip_font, vip_adj
The graphical parameters of the vips markings.
vip_legend_title
Character, the title of the vips
legend.
vip_legend_title_pos
A numeric vector of length
two, the xy position of the title within the vips
legend box.
vip_legend_title_cex, vip_legend_title_font, vip_legend_title_adj
The text parameters to be applied in the vips legend
title (par).
vip_legend_box_color
The background color of the
vips legend box.
vip_legend_key_cex
Numeric, the sizing factor of the keys
in the vips legend respect to the vips marking
in the plot.
vip_legend_key_margin
Numeric, the x position of the keys
within the vips legend box. Values from 0 to 1.
vip_legend_text_cex, vip_legend_text_font, vip_legend_text_adj
The text parameters of the text entries in the vips legend.
vip_legend_text_margin
Numeric, the x position of the
text entries in the vips legend box. Values from 0 to 1.
fitAnalysis_lwd, fitAnalysis_screePlot_color, fitAnalysis_stress_cex, fitAnalysis_stress_lab_cex, fitAnalysis_stress_axis_cex, fitAnalysis_stress_p_color, fitAnalysis_stress_l_color
The graphical parameters of the plot for fit analysis
correspondint to the ordination method (Scree plot for PCA, PCoA, LDA; Shepard or Stress plot for NMDS).
(par,
stressplot
of the vegan package).
test_text
A list of character vectors or expressions with the
lines of text presenting the results of statistical tests. A example
structure would be: list(c("first line", "second line"), "second paragraph"), bquote("third paragraph" ~ alpha == 2).
test_spacing_paragraph, test_spacing_line
Numeric, relative spacing
between paragraphs (list elements) and lines (character elements
within a list element, if more than one).
test_cex, test_font, test_adj
The parameters of
the text with the test results
(par).
fit_into_main
Logical, wheter to fit all elements
into the main plot. If TRUE, the 'fig' parameter of
every element is interpreted as relative to main_fig.
main_fig, group_legend_fig, vip_legend_fig, arrow_fig, fitAnalysis_fig, test_fig, x_title_fig, y_title_fig
The
fig parameter (par)
to place in the display region of the graphics device,
respectively, the main plot, the group and vip legends,
the fit analysis plot, the tests results, and the x
and y axes titles.
output_type
A character vector indicating the
output image types to be generated. Values accepted
are: "png", "tiff", "jpeg", "eps", and "preview"
(i.e. R graphics device).
open_new_device
Logical, wheter to build the plot in
a new graphics device. If FALSE, the biplot is
draw in the current device.
leave_device_open
Logical, wheter to leave the
graphics device open, e.g. to continue the plot
adding external elements. TRUE is only accepted if
output_type has a single value. If
output_type = c("preview"), the device is
left open by default.
directory
Character, the directory within the
working directory. For example, "MyFolder/".
file_name
Character, the name of the output file.
width, height
Numeric, the dimensions of the output image.
family
Character, the font family used in every text
in the plot, (par).
Details
This function allows customising virtually every
graphical parameter in a 2D biplot, including several extra
elements that may be useful for multivariate explorations.
It is focused mainly on improving basic visualization aspects
of ordination methods through 'classical' biplots. There
are several packages that address the creation of other
variations of biplot: BiplotGUI, GGEBiplotGUI, multibiplotGUI,
biplotbootGUI, NominalLogisticBiplot, OrdinalLogisticBiplot,
ade4, vegan, MultBiplotR.
When biplot_type = "default", the biplot processing
is done as in biplot.princomp, which
follows the definition of Gabriel (1971). As in this method,
when biplot_type = "pc.biplot", this function creates
biplots according with Gabriel and Odoroff (1990). Since there
are several types of biplot transformations,
it is possible to use 'scores' and 'loadings' that were
already transformed, passing biplot_type = NULL.
Groups can be represented as stars, ellipsoids, and/or colors,
using (s.class), which can be tracked by a
fully-customisable legend (group_legend arguments).
Individual observations deemed exceptional (vip = Very
Important Points) can be marked with custom symbols. Whenever
is are more than one type of marking (e.g. different
methods/criteria of outlier detection), the different symbols
can be presented in a legend (vip_legend arguments).
When desired, it is possible to display a Scree plot
representing the eigenvalues (Principal Components Analysis,
Principal Coordinates Analysis) or a Shepard or stress plot
(Nonmetric Multidimensional Scaling,
e.g. metaMDS in the vegan package) by
enabling show_fit_analysis. It is also possible to
display statistical test results by enabling show.tests
and introducing the corresponding lines in test_text.
The position of all elements (legend boxes, title and
subtitle, fit analysis plot and tests) can be customized
using the corresponding fig parameter.
Examples
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
## Not run:
# Use iris data
data("iris")
# get an ordination object
# ("PCA" is the default input of this function)
pca <- princomp(iris[, 1:4])
# Default plot using Species as the group factor
biplot_2d(pca, groups = iris$Species)
# Use the typical visualization,
# placing scores and loadings around the same origin
biplot_2d(pca, groups = iris$Species, detach_arrows = FALSE)
# Compare different versions of the classical biplot
# "default" vs. "pc.biplot"
biplot_2d(pca,
output_type = "preview",
leave_device_open = TRUE,
x_title = 'biplot_type = "default"',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0, 0.499, 0, 1))
biplot_2d(pca,
output_type = "preview",
open_new_device = FALSE,
biplot_type = "pc.biplot",
x_title = 'biplot_type = "pc.biplot"',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0.5099, 1, 0, 1))
# varying focus on representing distances
# between observations or between variables
biplot_2d(pca,
output_type = "png",
leave_device_open = TRUE,
rows_over_columns = 1,
x_title = 'observation-focused\nrows_over_columns = 1',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0, 0.329, 0, 1),
width = 1200)
biplot_2d(pca,
open_new_device = FALSE,
leave_device_open = TRUE,
rows_over_columns = 0.5,
x_title = 'compromise\nrows_over_columns = 0.5',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0.331, 0.659, 0, 1))
biplot_2d(pca,
open_new_device = FALSE,
rows_over_columns = 0,
biplot_type = "pc.biplot",
x_title = 'variable-focused\nrows_over_columns = 0',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0.661, 1, 0, 1))
# ---------------------------------------------------------
# Plot groups as different colors and point types (pch),
# make group star, ellipsis, and label invisible and
# add a group legend with a a custom title.
biplot_2d(pca,
groups = iris$Species,
group_color = NULL,
point_pch = c(1, 3, 2),
group_star_cex = 0,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = T,
group_legend_title = "Species")
# ---------------------------------------------------------
# Polish covariance arrows
# Abbreviate variables names
dimnames(pca$loadings)[[1]] <- c("SL", "SW", "PL", "PW")
# Set a specific justification (adj) for each variable label
arrow_label_adj_override <- rbind(c(-0.1, 0),
c(-0.1, 0.5),
c(0.5, 1.3),
c(0.5, 1.3))
row.names(arrow_label_adj_override) <-
dimnames(pca$loadings)[[1]]
# Plot: arrows with different colors and
# without the background grid
biplot_2d(pca,
groups = iris$Species,
point_pch = c(1, 3, 2),
group_star_cex = 0,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = T,
group_legend_title = "Species",
arrow_color = c("orange",
"blue",
"red",
"green"),
arrow_label_adj_override = arrow_label_adj_override,
show_grid = FALSE)
# ---------------------------------------------------------
# Get arbitrary Very Important Points
irisVIP <- list(setosa = (1:nrow(iris) == 16 |
1:nrow(iris) == 42),
versicolor=(1:nrow(iris) == 61),
virginica=(1:nrow(iris) == 107 |
1:nrow(iris) == 118 |
1:nrow(iris) == 132))
# Plot observations using their names and group by Species using only color.
# Mark the VIP and add the respective legend with custom characters.
biplot_2d(pca,
groups = iris$Species,
point_type = "label",
point_label = row.names(iris),
group_color = c("red", "blue", "green"),
group_star_cex = 0,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = TRUE,
group_legend_title = "",
vips = irisVIP,
vip_pch = c("X", "O", "+"),
vip_cex = c(2, 2, 3),
vip_legend_fig = c(0.01, 0.25, 0.7, 0.99),
show_axes = FALSE)
# ---------------------------------------------------------
# Test the setosa separation
irisDist <- dist(iris[, 1:4])
setosaSeparation <- iris$Species == "setosa"
## multivariate test for the setosa separation
require(vegan)
irisTests <- NULL
irisTests$permanova <- adonis(irisDist~setosaSeparation)
irisTests$permdisp2 <- permutest(betadisper(irisDist,setosaSeparation),
pairwise = TRUE)
# This function prepares a list of character vectors containing the test results
getTestText <- function(tests){
permanova_F <- as.character(round(tests$permanova$aov.tab$F.Model[1], 3))
permanova_pvalue <- as.character(round(tests$permanova$aov.tab$"Pr(>F)"[1], 3))
permanova_rSquared <- as.character(round(tests$permanova$aov.tab$R2[1], 3))
permdisp2_F <- as.character(round(tests$permdisp2$tab$F[1], 3))
permdisp2_pvalue <- as.character(round(tests$permdisp2$tab$"Pr(>F)"[1], 3))
text <- list(c(paste("PERMANOVA:\n F = ", permanova_F,
" (p = ", permanova_pvalue, ")\n", sep = ""),
c(expression(paste(" ", R^2, " =", sep = "")),
paste(" ", permanova_rSquared, sep = ""))),
paste("PERMDISP2:\n F = ", permdisp2_F,
" (p = ", permdisp2_pvalue,")", sep = ""))
return(text)
}
# Plot observations using points and groups as colored stars with no labels.
# Place tests results in the top left corner and give a custom horizontal title.
biplot_2d(pca,
groups = iris$Species,
group_color = NULL,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = TRUE,
group_legend_title = "",
test_text = resultText(irisTests),
test_fig = c(0.01,0.5,0.6,0.95),
show_axes = FALSE,
x_title = "testing setosa separation",
x_title_cex = 1.5)
## End(Not run)
Andros-Spica/biplot2d3d documentation built on June 10, 2020, 1:38 p.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.
Description Usage Arguments Details Examples
Description
Generates a 2D biplot representing the default
points/scores and loadings of an ordination object,
such as a PCA produced by princomp.
Usage
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 | biplot_2d(ordination_object, ordination_method = "PCA",
biplot_type = "default", rows_over_columns = 0.5, groups = NULL,
vips = NULL, detach_arrows = TRUE, show_grid = TRUE, show_axes = TRUE,
show_group_legend = FALSE, show_vip_legend = TRUE, show_arrows = TRUE,
show_fitAnalysis = TRUE, main_lwd = 2, grid_cex = 1, grid_font = 3,
grid_adj = 0.5, invert_coordinates = c(FALSE, FALSE), xlim = NULL,
ylim = NULL, x_title = "", y_title = "", x_title_cex = 1,
x_title_font = 2, y_title_cex = 1, y_title_font = 2, subtitle = NULL,
subtitle_position = "bottomleft", subtitle_cex = 1.2,
point_type = "point", point_pch = 1, point_size = 1,
point_label = NULL, point_label_cex = 1, point_label_font = 3,
point_label_adj = c(0.5, 0.5), point_label_adj_override = NULL,
arrow_color = "darkorange", arrow_mim_dist = 0, arrow_length = 0.2,
arrow_cex = 0.1, arrow_lwd = 2, arrow_label_cex = 1,
arrow_label_color = "black", arrow_label_font = 1,
arrow_label_adj = c(0.5, 0.5), arrow_label_adj_override = NULL,
group_color = "black", group_star_cex = 1, group_ellipse_cex = 1,
group_ellipse_axes = FALSE, group_label_cex = 1,
group_legend_title = "groups", group_legend_title_pos = c(0.5, 0.85),
group_legend_title_cex = 1, group_legend_title_font = 3,
group_legend_title_adj = 0.5, group_legend_box_color = "white",
group_legend_key_pch = 15, group_legend_key_cex = 1,
group_legend_key_lwd = 1, group_legend_key_margin = 0.15,
group_legend_text_margin = 0.25, group_legend_text_color = "black",
group_legend_text_cex = 1, group_legend_text_font = 1,
group_legend_text_adj = 0, vip_color = "black", vip_pch = c(0, 1, 5, 2,
6, 4, 3), vip_cex = c(5, 5, 5, 5, 5, 3, 3), vip_lwd = 5, vip_font = 1,
vip_adj = c(0.5, 0.5), vip_legend_title = "VIPs",
vip_legend_title_pos = c(0.5, 0.85), vip_legend_title_cex = 1,
vip_legend_title_font = 3, vip_legend_title_adj = 0.5,
vip_legend_box_color = "white", vip_legend_key_cex = 0.8,
vip_legend_key_margin = 0.15, vip_legend_text_cex = 1,
vip_legend_text_font = 1, vip_legend_text_adj = 0,
vip_legend_text_margin = 0.25, fitAnalysis_lwd = 3,
fitAnalysis_screePlot_color = c("grey", "white"),
fitAnalysis_stress_cex = 1, fitAnalysis_stress_lab_cex = 1,
fitAnalysis_stress_axis_cex = 1, fitAnalysis_stress_p_color = "darkgrey",
fitAnalysis_stress_l_color = "black", test_text = NULL,
test_spacing_paragraph = 0.8, test_spacing_line = 0.8, test_cex = 1,
test_font = 1, test_adj = 0.5, fit_into_main = FALSE, main_fig = c(0,
1, 0, 1), group_legend_fig = c(0.8, 0.99, 0.6, 0.9),
vip_legend_fig = c(0.78, 0.99, 0.3, 0.55), arrow_fig = c(0.69, 0.99, 0.01,
0.31), fitAnalysis_fig = c(0.02, 0.35, 0.06, 0.25), test_fig = c(0, 0.3,
0.8, 0.99), x_title_fig = c(0.25, 1, 0.85, 1), y_title_fig = c(0.91, 1, 0,
1), output_type = c("preview", "png"), open_new_device = TRUE,
leave_device_open = FALSE, directory = "", file_name = "2D Biplot",
width = 400, height = 400, family = "sans")
|
Arguments
ordination_object |
A R object containing a
direct and named reference to default ordination
outputs (i.e. |
ordination_method |
Character, the ordination method that was used to generate the ordination object: "PCA" for Principal Components Analysis (default), "PCoA" for Principal Coordinates Analysis, "NMDS" for Non-metric Multidimensional Scaling, and "LDA" for Linear Discriminant Analysis. |
biplot_type |
Character, indicating the type of
biplot transformation of data: "default" and
"pc.biplot", corresponding to the transformations
performed in |
rows_over_columns |
Numeric, the value defining the
degree in which distances between observations have
priority over distances between variables
(0 = observation-focused, 1 = variable-focused).
It corresponds to the argument |
groups |
A factor variable containing the group assignation of each point. |
vips |
A list of logical (Boolean) vectors identifying "Very Important Points" under different methods or criteria. |
detach_arrows |
Logical, wheter to display covariance arrows
a independent miniature plot, overlapping the main plot and
placed according to |
show_grid |
Logical, wheter to display the background grid. |
show_axes |
Logical, wheter to display the axes. |
show_group_legend |
Logical, whether to display a legend for groups. |
show_vip_legend |
Logical, whether to display a legend for vip criteria. |
show_arrows |
Logical, whether to show variable covariance arrows. |
show_fitAnalysis |
Logical, whether to display the fit analysis plot corresponding to the ordination method given (Scree plot or Shepard plot). |
main_lwd |
The line width to be used in the main plot.
( |
grid_cex, grid_font, grid_adj |
The scale, font type, and text justification of the grid notation. |
invert_coordinates |
Logical, vector of length two expressing which dimensions, if any, must be inverted before plotting (e.g. for aesthetical reasons). |
xlim, ylim |
the ranges to be encompassed by the x and
y axes, if NULL they are computed
(See |
x_title, y_title |
Character, texts to be placed in the x and y axes. |
x_title_cex, x_title_font |
The text parameters
of the x axis or horizontal title ( |
y_title_cex, y_title_font |
The text parameters
of the y axis or vertical title ( |
subtitle |
Character, a subtitle to be displayed
in the bottom-left corner of the plot
( |
subtitle_position |
Character, value indicating the
position of the subtitle in the main plot ("topleft",
"topright", "bottomleft", and "bottomright"; see
|
subtitle_cex |
the font size of the
subtitle ( |
point_type |
A character input accepting three
values: "point", "label" (displaying the content
of |
point_pch |
A number or a numerical vector given
to |
point_size |
The size or scale given to |
point_label |
A character vector labelling every
observation. It's length must be equal to the number
of rows in the points/scores of the ordination object
( |
point_label_cex, point_label_font, point_label_adj |
The text parameters of the points' labels
( |
point_label_adj_override |
A data frame with x,y
values to be passed to |
arrow_color |
The color or colors to be used in covariance arrows. |
arrow_mim_dist |
Numeric, the minimum distance of a arrow from the origin of arrows (i.e. zero covariance), in order for it to be displayed (range [0 = all arrows are displayed,1 = no arrow is displayed]). |
arrow_length |
Numeric, the scalar factor applied to loadings
to resize them respect to scores. If |
arrow_cex, arrow_lwd, arrow_label_cex, arrow_label_color, arrow_label_font, arrow_label_adj |
Graphical parameters of the covariance arrows and their
labels (see |
arrow_label_adj_override |
A data frame with x,y values
to be passed to |
group_color |
A vector containing the colors to
be used in each group (applied to points, labels,
stars and ellipses). If NULL, automatically assign
different colors from the |
group_star_cex, group_ellipse_cex, group_label_cex |
The
size or scale of the stars, ellipses, and labels
representing groups passed to |
group_ellipse_axes |
Logical, wheter to show
the ellipses axes, passed to |
group_legend_title |
Character, the title of the group legend. If equal NULL or "", no title is displayed. |
group_legend_title_pos |
A numeric vector of length two, the xy position of the title within the group legend box. |
group_legend_title_cex, group_legend_title_font, group_legend_title_adj |
The text parameters to be applied in the group legend
title ( |
group_legend_box_color |
The background color of the group legend box. |
group_legend_key_pch, group_legend_key_cex, group_legend_key_lwd |
The type, size and line width of the keys in the group legend. |
group_legend_key_margin |
The x position of the keys within the group legend box. Values from 0 to 1. |
group_legend_text_margin |
The x position of the text entries in the group legend. Values from 0 to 1. |
group_legend_text_color |
The color or colors of the text entries in the group legend. |
group_legend_text_cex, group_legend_text_font, group_legend_text_adj |
The text parameters of the text entries in the group legend. |
vip_color |
The color or colors to be used in vip markings. |
vip_pch |
A character vector containing the characters used for the vips markings under each criterion. |
vip_cex, vip_lwd, vip_font, vip_adj |
The graphical parameters of the vips markings. |
vip_legend_title |
Character, the title of the vips legend. |
vip_legend_title_pos |
A numeric vector of length two, the xy position of the title within the vips legend box. |
vip_legend_title_cex, vip_legend_title_font, vip_legend_title_adj |
The text parameters to be applied in the vips legend
title ( |
vip_legend_box_color |
The background color of the vips legend box. |
vip_legend_key_cex |
Numeric, the sizing factor of the keys in the vips legend respect to the vips marking in the plot. |
vip_legend_key_margin |
Numeric, the x position of the keys within the vips legend box. Values from 0 to 1. |
vip_legend_text_cex, vip_legend_text_font, vip_legend_text_adj |
The text parameters of the text entries in the vips legend. |
vip_legend_text_margin |
Numeric, the x position of the text entries in the vips legend box. Values from 0 to 1. |
fitAnalysis_lwd, fitAnalysis_screePlot_color, fitAnalysis_stress_cex, fitAnalysis_stress_lab_cex, fitAnalysis_stress_axis_cex, fitAnalysis_stress_p_color, fitAnalysis_stress_l_color |
The graphical parameters of the plot for fit analysis
correspondint to the ordination method (Scree plot for PCA, PCoA, LDA; Shepard or Stress plot for NMDS).
( |
test_text |
A list of character vectors or expressions with the
lines of text presenting the results of statistical tests. A example
structure would be: |
test_spacing_paragraph, test_spacing_line |
Numeric, relative spacing between paragraphs (list elements) and lines (character elements within a list element, if more than one). |
test_cex, test_font, test_adj |
The parameters of
the text with the test results
( |
fit_into_main |
Logical, wheter to fit all elements
into the main plot. If TRUE, the 'fig' parameter of
every element is interpreted as relative to |
main_fig, group_legend_fig, vip_legend_fig, arrow_fig, fitAnalysis_fig, test_fig, x_title_fig, y_title_fig |
The
|
output_type |
A character vector indicating the output image types to be generated. Values accepted are: "png", "tiff", "jpeg", "eps", and "preview" (i.e. R graphics device). |
open_new_device |
Logical, wheter to build the plot in a new graphics device. If FALSE, the biplot is draw in the current device. |
leave_device_open |
Logical, wheter to leave the
graphics device open, e.g. to continue the plot
adding external elements. TRUE is only accepted if
|
directory |
Character, the directory within the working directory. For example, "MyFolder/". |
file_name |
Character, the name of the output file. |
width, height |
Numeric, the dimensions of the output image. |
family |
Character, the font family used in every text
in the plot, ( |
Details
This function allows customising virtually every
graphical parameter in a 2D biplot, including several extra
elements that may be useful for multivariate explorations.
It is focused mainly on improving basic visualization aspects
of ordination methods through 'classical' biplots. There
are several packages that address the creation of other
variations of biplot: BiplotGUI, GGEBiplotGUI, multibiplotGUI,
biplotbootGUI, NominalLogisticBiplot, OrdinalLogisticBiplot,
ade4, vegan, MultBiplotR.
When biplot_type = "default", the biplot processing
is done as in biplot.princomp, which
follows the definition of Gabriel (1971). As in this method,
when biplot_type = "pc.biplot", this function creates
biplots according with Gabriel and Odoroff (1990). Since there
are several types of biplot transformations,
it is possible to use 'scores' and 'loadings' that were
already transformed, passing biplot_type = NULL.
Groups can be represented as stars, ellipsoids, and/or colors,
using (s.class), which can be tracked by a
fully-customisable legend (group_legend arguments).
Individual observations deemed exceptional (vip = Very
Important Points) can be marked with custom symbols. Whenever
is are more than one type of marking (e.g. different
methods/criteria of outlier detection), the different symbols
can be presented in a legend (vip_legend arguments).
When desired, it is possible to display a Scree plot
representing the eigenvalues (Principal Components Analysis,
Principal Coordinates Analysis) or a Shepard or stress plot
(Nonmetric Multidimensional Scaling,
e.g. metaMDS in the vegan package) by
enabling show_fit_analysis. It is also possible to
display statistical test results by enabling show.tests
and introducing the corresponding lines in test_text.
The position of all elements (legend boxes, title and
subtitle, fit analysis plot and tests) can be customized
using the corresponding fig parameter.
Examples
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 | ## Not run:
# Use iris data
data("iris")
# get an ordination object
# ("PCA" is the default input of this function)
pca <- princomp(iris[, 1:4])
# Default plot using Species as the group factor
biplot_2d(pca, groups = iris$Species)
# Use the typical visualization,
# placing scores and loadings around the same origin
biplot_2d(pca, groups = iris$Species, detach_arrows = FALSE)
# Compare different versions of the classical biplot
# "default" vs. "pc.biplot"
biplot_2d(pca,
output_type = "preview",
leave_device_open = TRUE,
x_title = 'biplot_type = "default"',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0, 0.499, 0, 1))
biplot_2d(pca,
output_type = "preview",
open_new_device = FALSE,
biplot_type = "pc.biplot",
x_title = 'biplot_type = "pc.biplot"',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0.5099, 1, 0, 1))
# varying focus on representing distances
# between observations or between variables
biplot_2d(pca,
output_type = "png",
leave_device_open = TRUE,
rows_over_columns = 1,
x_title = 'observation-focused\nrows_over_columns = 1',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0, 0.329, 0, 1),
width = 1200)
biplot_2d(pca,
open_new_device = FALSE,
leave_device_open = TRUE,
rows_over_columns = 0.5,
x_title = 'compromise\nrows_over_columns = 0.5',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0.331, 0.659, 0, 1))
biplot_2d(pca,
open_new_device = FALSE,
rows_over_columns = 0,
biplot_type = "pc.biplot",
x_title = 'variable-focused\nrows_over_columns = 0',
x_title_fig = c(0, 1, 0.9, 1),
fit_into_main = TRUE,
main_fig = c(0.661, 1, 0, 1))
# ---------------------------------------------------------
# Plot groups as different colors and point types (pch),
# make group star, ellipsis, and label invisible and
# add a group legend with a a custom title.
biplot_2d(pca,
groups = iris$Species,
group_color = NULL,
point_pch = c(1, 3, 2),
group_star_cex = 0,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = T,
group_legend_title = "Species")
# ---------------------------------------------------------
# Polish covariance arrows
# Abbreviate variables names
dimnames(pca$loadings)[[1]] <- c("SL", "SW", "PL", "PW")
# Set a specific justification (adj) for each variable label
arrow_label_adj_override <- rbind(c(-0.1, 0),
c(-0.1, 0.5),
c(0.5, 1.3),
c(0.5, 1.3))
row.names(arrow_label_adj_override) <-
dimnames(pca$loadings)[[1]]
# Plot: arrows with different colors and
# without the background grid
biplot_2d(pca,
groups = iris$Species,
point_pch = c(1, 3, 2),
group_star_cex = 0,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = T,
group_legend_title = "Species",
arrow_color = c("orange",
"blue",
"red",
"green"),
arrow_label_adj_override = arrow_label_adj_override,
show_grid = FALSE)
# ---------------------------------------------------------
# Get arbitrary Very Important Points
irisVIP <- list(setosa = (1:nrow(iris) == 16 |
1:nrow(iris) == 42),
versicolor=(1:nrow(iris) == 61),
virginica=(1:nrow(iris) == 107 |
1:nrow(iris) == 118 |
1:nrow(iris) == 132))
# Plot observations using their names and group by Species using only color.
# Mark the VIP and add the respective legend with custom characters.
biplot_2d(pca,
groups = iris$Species,
point_type = "label",
point_label = row.names(iris),
group_color = c("red", "blue", "green"),
group_star_cex = 0,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = TRUE,
group_legend_title = "",
vips = irisVIP,
vip_pch = c("X", "O", "+"),
vip_cex = c(2, 2, 3),
vip_legend_fig = c(0.01, 0.25, 0.7, 0.99),
show_axes = FALSE)
# ---------------------------------------------------------
# Test the setosa separation
irisDist <- dist(iris[, 1:4])
setosaSeparation <- iris$Species == "setosa"
## multivariate test for the setosa separation
require(vegan)
irisTests <- NULL
irisTests$permanova <- adonis(irisDist~setosaSeparation)
irisTests$permdisp2 <- permutest(betadisper(irisDist,setosaSeparation),
pairwise = TRUE)
# This function prepares a list of character vectors containing the test results
getTestText <- function(tests){
permanova_F <- as.character(round(tests$permanova$aov.tab$F.Model[1], 3))
permanova_pvalue <- as.character(round(tests$permanova$aov.tab$"Pr(>F)"[1], 3))
permanova_rSquared <- as.character(round(tests$permanova$aov.tab$R2[1], 3))
permdisp2_F <- as.character(round(tests$permdisp2$tab$F[1], 3))
permdisp2_pvalue <- as.character(round(tests$permdisp2$tab$"Pr(>F)"[1], 3))
text <- list(c(paste("PERMANOVA:\n F = ", permanova_F,
" (p = ", permanova_pvalue, ")\n", sep = ""),
c(expression(paste(" ", R^2, " =", sep = "")),
paste(" ", permanova_rSquared, sep = ""))),
paste("PERMDISP2:\n F = ", permdisp2_F,
" (p = ", permdisp2_pvalue,")", sep = ""))
return(text)
}
# Plot observations using points and groups as colored stars with no labels.
# Place tests results in the top left corner and give a custom horizontal title.
biplot_2d(pca,
groups = iris$Species,
group_color = NULL,
group_ellipse_cex = 0,
group_label_cex = 0,
show_group_legend = TRUE,
group_legend_title = "",
test_text = resultText(irisTests),
test_fig = c(0.01,0.5,0.6,0.95),
show_axes = FALSE,
x_title = "testing setosa separation",
x_title_cex = 1.5)
## End(Not run)
|
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.