Could using grid.arrange affect plot output in R? - r

Question

For a function that accepts a data frame, a plot switch value, and bin size vector, I'm trying to create plots for each count-bin and each density-bin size combination for every numerical variable column when a plot switch value of 'grid' is inputted. I have it working that I get individual plots for each combination, but when I tried to put the plots for each numerical variable on a grid and I'm seeing variations in the actual plots even though as far as I can tell I didn't alter the calculations nor the plot commands.
For instance, notice that, even accounting for scale variations, y-density plots are different in the images below. Tried looking at the documentation for grid.arrange and list just to make sure I wasn't missing something, but I still can't figure it out. Did I somehow change the plot command without realizing it?
Tested with:
data(diamonds, package = "ggplot2")
test <- diamonds[1:100,]
hstPlts(test,ps='grid') #without grid
hst1(test,ps='grid') #with grid
hstPlts <- function(df,ps = 'off',bnSize=c(30)){
#hstPlts() accepts any data frame and if the ps parameter is “on” or
#“grid”, then plot a pair of blue histograms with a vertical red line at the
#mean
#for every numerical variable at each number of bins integer specified in
#the bin
#vector parameter. if the plot switch is set to “grid”, there should be a
#grid for each count-bin combination and a separate grid for
#each density-bin size combination.
#parameters:
#df : data frame
#ps : plot switch defaulted to 'off' with two additional value options -
#'on' and 'grid'
#bnSize : vector containing numeric values representing number of bins for
histograms
num_var <- df[sapply(df,is.numeric)] #extract numeric columns
for(i in 1:length(bnSize)){ #iterate through each bin size
for(j in 1:(ncol(num_var))){ #iterate through each numeric variable
if(ps=='on' | ps =='grid'){ #conditional for both 'on' and 'grid'
#values
bnWid <- (max(num_var[,j]) - min(num_var[,j]))/bnSize[i] # compute
#bin widths
vrMn <- mean(num_var[[j]]) #compute column means for red line
mean = sprintf("%8.3f ", vrMn) #set up label for mean line with
#formatted decimals
cntPlt <- ggplot(num_var, aes(x=num_var[,j])) + #plot count
#histogram with numeric variable on x-axis
geom_histogram(colour = "blue", fill = "blue", binwidth =
bnWid) +
#detail bar fill colors and histogram bin widths
geom_vline(xintercept = mean(num_var[,j]),colour="red") +
#mean line
abs(x=colnames(num_var)[j]) #label x-axis
densPlt <- cntPlt + aes(y = ..density..) + labs(y = "density")
#create corresponding density histogram
print(cntPlt)
print(densPlt)
}
if(ps == 'grid'){ #conditional for just 'grid' value
grdPlt <- ggplot(num_var, aes(x=num_var[,j])) + #create plot for
#each count-bin combination and a separate grid for
#each density-bin size combination
geom_histogram(colour = "blue", fill = "blue", binwidth =
bnWid) +
#detail bar fill colors and histogram bin widths
labs(x=colnames(num_var)[j]) #label x-axis
print(grdPlt)
print(grdPlt + aes(y = ..density..) + labs(y = "density"))
}
}
}
}
hst1 <- function(df,ps = 'off',bsizes=c(30)){
numvar <- df[sapply(df,is.numeric)]
if(ps=='on')
for(i in 1:length(bsizes)){
for(j in 1:(ncol(numvar))){
bwidth <- (max(numvar[,j]) - min(numvar[,j]))/bsizes[i]
var_mean <- mean(numvar[[j]])
mean = sprintf("%8.3f ", var_mean)
counts <- ggplot(numvar, aes(x=numvar[,j])) +
geom_histogram(colour = "blue", fill = "blue", binwidth
= bwidth) +
geom_vline(xintercept = mean(numvar[,j]),colour="red") +
labs(x=colnames(numvar)[j]) # Labeling the x axis
dense <- counts + aes(y = ..density..) + labs(y = "density")
print(counts)
print(dense)
}
}
}
else if(ps == 'grid'){
for(i in 1:length(bsizes)){
cntLst <- list()
denseLst <- list()
for(j in 1:(ncol(numvar))){
bwidth <- (max(numvar[,j]) - min(numvar[,j]))/bsizes[i]
cntGrPlt <- ggplot(numvar, aes(x=numvar[,j])) +
geom_histogram(colour = "blue", fill = "blue", binwidth =
bwidth) +
labs(x=colnames(numvar)[j])
cntLst[[length(cntLst)+1]] <- cntGrPlt
denseLst[[j]] <- cntGrPlt + aes(y = ..density..) + labs(y =
"density")
}
grid.arrange(grobs=cntLst,ncol=2)
grid.arrange(grobs=denseLst,ncol=2)
}
}
}