In jzemmels/impressions: Diagnose Cartridge Case Comparison Algorithms
knitr::opts_chunk$set(
collapse = TRUE,
fig.align = "center",
out.width = "60%",
comment = "#>"
)
The x3pPlot() (or x3p_plot()) function creates a visualization of one or more x3p objects using the ggplot2 package.
We use base ggplot2 in x3pPlot(), so you can easily change these settings if you're already familiar.
The x3pPlot() function essentially consists of a call to geom_raster() with possible faceting and setting default features such as the pixel color scale, legend, etc.
In this vignette, we discuss how to change characteristics of the plot.
Basic usage of x3pPlot
library(impressions)
library(ggplot2)
data("K013sA1")
data("K013sA2")
First, let's make sure that we understand the basic functionality of x3pPlot().
This section will discuss how the arguments of x3pPlot() expose some of the underlying ggplot2 functionality to change the output.
Experts of ggplot2 can ignore these arguments by simply adding (+) desired changes to the ggplot2 object returned by x3pPlot().
The sections after this one show how you can extend the functionality of x3pPlot() using ggplot2.
The x3pPlot function takes as its first argument(s) a sequence or list of x3p objects.
The code below demonstrates what we mean by a "sequence" in that they can be separated by commas.
We see that the output visualizes the surface values of the two x3p objects on the same color scale using faceting.
If not specified, the facet labels are set to x3p[0-9]{1,} up to the number of x3p objects passed.
x3pPlot(K013sA1,K013sA2)
Wrapping these objects within a list() provides the exact same output.^[We provide this flexibility of input because of the two expected use cases.
One common use case will be that you have an x3p object loaded directly into your environment (e.g., by running data("K013sA1") as in this vignette).
In this case, it's nicer to pass the name of the object directly into the x3pPlot() function rather than wrapping it first in a list.
The second expected use case is when you have a list of x3p objects stored in the column of a data frame-like object such as a tibble() -- this is the case when comparing cartridge cases using the cmcR or scored packages.
In this case, it's easier to pass the whole column, like df$x3pCol, than flattening the column first.]
# exact same output as above, so not shown:
# x3pPlot(list(K013sA1, K013sA2))
x3p.names argument
The other arguments in x3pPlot() change characteristics of the plot.
First, we can change the facet labels; either by setting the x3p.names object or by passing a named list.
Note that x3p.names is ignored if a named list is used.
x3pPlot(K013sA1,K013sA2,x3p.names = c("Scan A","Scan B"))
# exact same output as above, so not shown:
# x3pPlot(list("Scan A" = K013sA1, "Scan B" = K013sA2))
output argument
The output argument specifies whether the x3p objects are visualized on the same color scale using faceting, or if each x3p object is visualized on its own color scale and the output is a list.
The former is useful for comparing the surface values across two or more scans.
The latter output is useful if you have x3p objects whose surface values are on different scales, meaning one scan's surface values would "drown out" others.
x3pPlot(K013sA1,K013sA2,output = "list")
height.colors, height.quantiles, and na.value arguments
By default, the color scheme used in x3pPlot() is a mapping of the x3p object surface value quantiles to a divergent purple, white, orange color scale.
The median surface value is mapped to white (technically a very light gray) while values below/above the median are mapped to progressively darker shades of purple/orange, respectively.
The na.value argument is set to "gray65" by default.
The image below summarizes the mapping between 11 surface quantiles and hexidecimal colors.
We chose this mapping because we found that it highlighted the distinguishable markings on a cartridge case surface well.
{width=100% align="center"}
The height.colors and height.quantiles arguments allow you to change the color scale and quantile mapping.
For example, the code below maps equally-spaced quantiles to grayscale.
This color scale doesn't distinguish between values below or above the median surface value.
We also change the na.value argument for illustration.
x3pPlot(K013sA1,K013sA2,
height.colors = paste0("gray",seq(0,100,by = 10)),
height.quantiles = seq(0,1,by = .1),
na.value = "red")
legend.quantiles and legend.length arguments
Keen eyes may be twitching at the plot legends in the previous examples.
The tick marks on the plot legend are spaced such that the first, second, and third quartile labels are bunched-up and overlapping.
We can give the legend quantile labels some breathing room using the legend.quantiles argument.
We can also adjust the legend length using the legend.length argument, which internally is passed to the barheight argument of guide_colorbar().
x3pPlot(K013sA1,K013sA2,
legend.quantiles = c(0,.01,.5,.99,1),
legend.length = grid::unit(1,"in"))
Color scale
We've already discussed how you can redefine the quantile color scale mapping in the x3pPlot() function.
If you want to change to a color scale that isn't based on quantiles, then you can simply overwrite the old color scale using your choice of scale_fill_* function.
The scale_fill_gradient2() example below illustrates why the previous quantile mapping is useful -- most of the surface values aren't extreme or distinguishable, so the quantile mapping makes the less populous, more interesting impressions more visible.
Note that you will get a warning for overwriting the previous color scale that you can suppress by wrapping the visualization call as suppressWarning(print(...))
plt <- x3pPlot(K013sA1,K013sA2,legend.length = grid::unit(1,"in"))
plt +
scale_fill_gradient2(low = "purple",mid = "white",high = "darkorange",midpoint = 0)
Legend
The legend can be changed using theme() and/or guide_colorbar() as below.
plt +
theme(legend.position = "bottom") +
guides(fill = guide_colourbar(barwidth = grid::unit(2,"in"),
barheight = grid::unit(.1,"in"),
title = "New Title",
title.position = "bottom",
label.position = "top",
label.theme = element_text(size = 6,
angle = 90),
title.theme = element_text(size = 8),
frame.colour = "white",
ticks.colour = "white"))
Axis Features
We hide the axis information by default.
However you might want to include axis text or ticks.
You can add a new theme to the plot to override the previous settings.
We prefer theme_minimal(), but you do you.
plt +
theme_minimal()
Note that the axis text below shows pixel-wise increments by default.
If you instead want to show the physical size of each scan, say in micrometers ("microns" or $\mu$m for short), then you'll need information about the scan resolution.
In this example, both scans K013sA1 and K013sA2 have a lateral resolution of approximately $7.4 \times 10^{-6}$ meters, or 7.4 micrometers, per pixel as you can see in the incrementY (or incrementX) element of the x3p objects.
We use scale_x_continuous() and scale_y_continuous() to change the axis labels to the micron scale.
Notice how we use the expression() function for the $\mu$m abbreviation.
We see that both scans are approximately 3000 $\mu$m in both directions.
K013sA1$header.info$incrementY #same as incrementX
# K013sA2$header.info$incrementY # same as above
plt +
theme_minimal() +
scale_x_continuous(labels = function(x) round(x*K013sA1$header.info$incrementX*1e6)) +
scale_y_continuous(labels = function(y) round(y*K013sA1$header.info$incrementY*1e6)) +
labs(x = expression("Horizontal Extent ["*mu*"m]"),
y = expression("Vertical Extent ["*mu*"m]"))
Faceting
Faceting allows multiple x3p objects to be visualized on the same color scale, which makes it easier to compare their surface values.
You can change the nature of the faceting by adding facet_wrap() or facet_grid(), which will overwrite the previous settings.
Within the x3pPlot() function, we convert each x3p object into a data frame and add a column called x3p that tracks the label associated with each scan.
You can facet the plot on this x3p column.
In the example below, you can see how the output changes depending on the faceting function used.
plt +
facet_wrap(~x3p,ncol = 1)
plt +
facet_grid(rows = vars(x3p))
We can also change the facet labeling using labeller() as in the example below.
plt +
facet_wrap(~ x3p,
labeller = labeller(x3p = c(x3p1 = "Scan A",
x3p2 = "Scan B"),
.default = function(string) toupper(string)))
jzemmels/impressions documentation built on Nov. 18, 2024, 10:30 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.
knitr::opts_chunk$set( collapse = TRUE, fig.align = "center", out.width = "60%", comment = "#>" )
The x3pPlot() (or x3p_plot()) function creates a visualization of one or more x3p objects using the ggplot2 package.
We use base ggplot2 in x3pPlot(), so you can easily change these settings if you're already familiar.
The x3pPlot() function essentially consists of a call to geom_raster() with possible faceting and setting default features such as the pixel color scale, legend, etc.
In this vignette, we discuss how to change characteristics of the plot.
Basic usage of x3pPlot
library(impressions) library(ggplot2) data("K013sA1") data("K013sA2")
First, let's make sure that we understand the basic functionality of x3pPlot().
This section will discuss how the arguments of x3pPlot() expose some of the underlying ggplot2 functionality to change the output.
Experts of ggplot2 can ignore these arguments by simply adding (+) desired changes to the ggplot2 object returned by x3pPlot().
The sections after this one show how you can extend the functionality of x3pPlot() using ggplot2.
The x3pPlot function takes as its first argument(s) a sequence or list of x3p objects.
The code below demonstrates what we mean by a "sequence" in that they can be separated by commas.
We see that the output visualizes the surface values of the two x3p objects on the same color scale using faceting.
If not specified, the facet labels are set to x3p[0-9]{1,} up to the number of x3p objects passed.
x3pPlot(K013sA1,K013sA2)
Wrapping these objects within a list() provides the exact same output.^[We provide this flexibility of input because of the two expected use cases.
One common use case will be that you have an x3p object loaded directly into your environment (e.g., by running data("K013sA1") as in this vignette).
In this case, it's nicer to pass the name of the object directly into the x3pPlot() function rather than wrapping it first in a list.
The second expected use case is when you have a list of x3p objects stored in the column of a data frame-like object such as a tibble() -- this is the case when comparing cartridge cases using the cmcR or scored packages.
In this case, it's easier to pass the whole column, like df$x3pCol, than flattening the column first.]
# exact same output as above, so not shown: # x3pPlot(list(K013sA1, K013sA2))
x3p.names argument
The other arguments in x3pPlot() change characteristics of the plot.
First, we can change the facet labels; either by setting the x3p.names object or by passing a named list.
Note that x3p.names is ignored if a named list is used.
x3pPlot(K013sA1,K013sA2,x3p.names = c("Scan A","Scan B")) # exact same output as above, so not shown: # x3pPlot(list("Scan A" = K013sA1, "Scan B" = K013sA2))
output argument
The output argument specifies whether the x3p objects are visualized on the same color scale using faceting, or if each x3p object is visualized on its own color scale and the output is a list.
The former is useful for comparing the surface values across two or more scans.
The latter output is useful if you have x3p objects whose surface values are on different scales, meaning one scan's surface values would "drown out" others.
x3pPlot(K013sA1,K013sA2,output = "list")
height.colors, height.quantiles, and na.value arguments
By default, the color scheme used in x3pPlot() is a mapping of the x3p object surface value quantiles to a divergent purple, white, orange color scale.
The median surface value is mapped to white (technically a very light gray) while values below/above the median are mapped to progressively darker shades of purple/orange, respectively.
The na.value argument is set to "gray65" by default.
The image below summarizes the mapping between 11 surface quantiles and hexidecimal colors.
We chose this mapping because we found that it highlighted the distinguishable markings on a cartridge case surface well.
{width=100% align="center"}
The height.colors and height.quantiles arguments allow you to change the color scale and quantile mapping.
For example, the code below maps equally-spaced quantiles to grayscale.
This color scale doesn't distinguish between values below or above the median surface value.
We also change the na.value argument for illustration.
x3pPlot(K013sA1,K013sA2, height.colors = paste0("gray",seq(0,100,by = 10)), height.quantiles = seq(0,1,by = .1), na.value = "red")
legend.quantiles and legend.length arguments
Keen eyes may be twitching at the plot legends in the previous examples.
The tick marks on the plot legend are spaced such that the first, second, and third quartile labels are bunched-up and overlapping.
We can give the legend quantile labels some breathing room using the legend.quantiles argument.
We can also adjust the legend length using the legend.length argument, which internally is passed to the barheight argument of guide_colorbar().
x3pPlot(K013sA1,K013sA2, legend.quantiles = c(0,.01,.5,.99,1), legend.length = grid::unit(1,"in"))
Color scale
We've already discussed how you can redefine the quantile color scale mapping in the x3pPlot() function.
If you want to change to a color scale that isn't based on quantiles, then you can simply overwrite the old color scale using your choice of scale_fill_* function.
The scale_fill_gradient2() example below illustrates why the previous quantile mapping is useful -- most of the surface values aren't extreme or distinguishable, so the quantile mapping makes the less populous, more interesting impressions more visible.
Note that you will get a warning for overwriting the previous color scale that you can suppress by wrapping the visualization call as suppressWarning(print(...))
plt <- x3pPlot(K013sA1,K013sA2,legend.length = grid::unit(1,"in")) plt + scale_fill_gradient2(low = "purple",mid = "white",high = "darkorange",midpoint = 0)
Legend
The legend can be changed using theme() and/or guide_colorbar() as below.
plt + theme(legend.position = "bottom") + guides(fill = guide_colourbar(barwidth = grid::unit(2,"in"), barheight = grid::unit(.1,"in"), title = "New Title", title.position = "bottom", label.position = "top", label.theme = element_text(size = 6, angle = 90), title.theme = element_text(size = 8), frame.colour = "white", ticks.colour = "white"))
Axis Features
We hide the axis information by default.
However you might want to include axis text or ticks.
You can add a new theme to the plot to override the previous settings.
We prefer theme_minimal(), but you do you.
plt + theme_minimal()
Note that the axis text below shows pixel-wise increments by default.
If you instead want to show the physical size of each scan, say in micrometers ("microns" or $\mu$m for short), then you'll need information about the scan resolution.
In this example, both scans K013sA1 and K013sA2 have a lateral resolution of approximately $7.4 \times 10^{-6}$ meters, or 7.4 micrometers, per pixel as you can see in the incrementY (or incrementX) element of the x3p objects.
We use scale_x_continuous() and scale_y_continuous() to change the axis labels to the micron scale.
Notice how we use the expression() function for the $\mu$m abbreviation.
We see that both scans are approximately 3000 $\mu$m in both directions.
K013sA1$header.info$incrementY #same as incrementX # K013sA2$header.info$incrementY # same as above plt + theme_minimal() + scale_x_continuous(labels = function(x) round(x*K013sA1$header.info$incrementX*1e6)) + scale_y_continuous(labels = function(y) round(y*K013sA1$header.info$incrementY*1e6)) + labs(x = expression("Horizontal Extent ["*mu*"m]"), y = expression("Vertical Extent ["*mu*"m]"))
Faceting
Faceting allows multiple x3p objects to be visualized on the same color scale, which makes it easier to compare their surface values.
You can change the nature of the faceting by adding facet_wrap() or facet_grid(), which will overwrite the previous settings.
Within the x3pPlot() function, we convert each x3p object into a data frame and add a column called x3p that tracks the label associated with each scan.
You can facet the plot on this x3p column.
In the example below, you can see how the output changes depending on the faceting function used.
plt + facet_wrap(~x3p,ncol = 1) plt + facet_grid(rows = vars(x3p))
We can also change the facet labeling using labeller() as in the example below.
plt + facet_wrap(~ x3p, labeller = labeller(x3p = c(x3p1 = "Scan A", x3p2 = "Scan B"), .default = function(string) toupper(string)))
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.