I am modifying this tutorial to produce a ggplot of polar vectors. Within my own data, and absent from the tutorial, are test groups by which I would like to use to facet my plot (hist.by.groups_deg$Testnames). When I facet the plot by the variable intended in the tutorial (comparison), each facet has only one polar vector, and the y-axis appears to be scaled correctly. However, when I switch to Testnames and try to plot more than one vector on the same facet, I get the following error:
Warning message:
Removed 6 rows containing missing values (geom_bar()).
The tutorial code standardizes the y-axis histogram counts by dividing those counts by the highest number of counts in the histogram, such that the data range for the y-axis should be 0 to 1, which appears to be the case when I check the dataframe before plotting. However, after receiving the warning message, I increased the maximum value on my y-axis and several were suddenly exceeding 1. Can someone help me properly standardize my y-axis so I can compare my groups on the same facet?
rm(list=ls())
graphics.off()
set.seed(1)
library(SIBER)
library(ggplot2)
library(magrittr) # to enable piping
library(dplyr)
# load in the included demonstration dataset
data("demo.siber.data")
# create the siber object
siber.example <- createSiberObject(demo.siber.data)
# options for running jags
parms <- list()
parms$n.iter <- 2 * 10^4 # number of iterations to run the model for
parms$n.burnin <- 1 * 10^3 # discard the first set of values
parms$n.thin <- 10 # thin the posterior by this many
parms$n.chains <- 2 # run this many chains
parms$save.output = FALSE
parms$save.dir = tempdir()
# define the priors
priors <- list()
priors$R <- 1 * diag(2)
priors$k <- 2
priors$tau.mu <- 1.0E-3
# fit the ellipses which uses an Inverse Wishart prior
# on the covariance matrix Sigma, and a vague normal prior on the
# means. Fitting is via the JAGS method.
ellipses.posterior <- siberMVN(siber.example, parms, priors)
# extract the centroids from the fitted model object
centroids <- siberCentroids(ellipses.posterior)
# calculate pairwise polar vectors among all groups
angles_distances <- allCentroidVectors(centroids, do.plot = FALSE)
my.hist <- function(df){
test <- hist(df$angles,
breaks = seq(from = -pi, to = pi, length = 60),
plot = FALSE)
X <- data.frame(counts = test$counts, mids = test$mids, dens = test$density,
counts.stdzd = test$counts / max(test$counts))
return(X)
}
# calculate the points for each group's ellipse
hist.by.groups <- angles_distances %>% group_by(comparison) %>%
do(my.hist(.))
### HERE IS WHERE I START MODIFYING THINGS ###
# Radians stress me out so I'm going to try converting them to degrees
hist.by.groups_deg <- hist.by.groups %>%
mutate(Degrees = mids*(180/pi))
# Creating test names to facet by
hist.by.groups_deg$TestName = c(rep("A", 125),
rep("B", 125),
rep("C", 635))
# Generating polar histograms
cols <- c("#0570b0", "#8c96c6", "#74a9cf", "#8856a7", "#d7b5d8", "#b3cde3", "#810f7c", "#66c2a4", "#bdc9e1", "#df65b0","salmon","pink","gray","beige","tan")
all.roses <- ggplot(data = hist.by.groups_deg, aes(x = Degrees, y = counts.stdzd, color = comparison, fill = comparison)) +
# Default here is stat = "count", but we are providing the y values so we override with stat = "identity"
geom_hline(yintercept = seq(0, 1, by = 0.25), colour = "grey90", linewidth = 0.2) +
geom_vline(xintercept = seq(0, 180, by = 45), colour = "grey90", linewidth = 0.2) +
geom_vline(xintercept = seq(0, -180, by = -45), colour = "grey90", linewidth = 0.2) +
geom_bar(stat = "identity", alpha = .7) +
scale_x_continuous(
limits = c(-180,180),
breaks = c(-180, -90, 0, 90, 180),
labels = c("","-90","0","90", "")) +
scale_y_continuous(
limits = c(0,1)) +
# Even though we've changed our units to degrees, start argument is in radians
coord_polar(start = pi/2, direction = -1) +
facet_wrap( ~ TestName, nrow = 2) +
scale_fill_manual(
values = cols) +
scale_color_manual(
values = cols) +
theme(axis.ticks.y = element_blank(),
axis.text.y = element_blank(),
legend.position = "none",
panel.border = element_blank(),
panel.grid = element_blank()) +
labs(y = "Counts (standardized by the proportion of total values)",
x = expression("Angle between ellipse centroids (\u00B0)"))
print(all.roses)
This turned out to be a problem with
geom_barrather than a problem withSIBERorcoord_polar. Despite having standardized my own y-axis in the input dataframe, this was overridden by the defaultpositionargument in geom_bar ("stacked"), so that when I added several bar plots to the same facet, any overlapping bars were stacked on top of each other, sometimes doubling or tripling the scale of the y axis. Specifying position = "identity" allows the bar charts to overlap without stacking: