Save multiple ggplots with different layout matrices - r

Question

Currently I'm creating multiple plots with regional data and save them to a PDF file. This works without problems, thanks to an SO post I've found (use grid.arrange over multiple pages or marrangeGrob with a layout_matrix).
This is my code so far:
library(ggplot2)
library(gridExtra)
library(dplyr)
data <- data.frame(
region = c("region 1", "region 2", "region 3", rep("region 4", 2), rep("region 5", 2)),
countries = c("country 1", "country 2", "country 3", "country 4", "country 5", "country 6", "country 7"),
dummydata1 = c(rep(1, 7)),
dummydata2 = c(rep(2, 7))
)
criterias <- list()
criterias[[ 'region_1' ]] <- data %>% filter(region == 'region 1')
criterias[[ 'region_2' ]] <- data %>% filter(region == 'region 2')
criterias[[ 'region_3' ]] <- data %>% filter(region == 'region 3')
criterias[[ 'region_4' ]] <- data %>% filter(region == 'region 4')
criterias[[ 'region_5' ]] <- data %>% filter(region == 'region 5')
# This layout matrix should be used for the regional plots
# Don't wonder about the strange numbering, some plots came later
# and it was easier to modify the matrix then all other functions.
regionLayout <- rbind(
c(1,1,1,1,1,2),
c(NULL,NULL,3,3,NULL,NULL),
c(9,9,4,4,10,10),
c(6,6,6,7,7,7),
c(6,6,6,7,7,7),
c(6,6,6,7,7,7),
c(6,6,6,7,7,7),
c(6,6,6,7,7,7),
c(6,6,6,7,7,7)
)
# This is just a dummy function
# The actual function creates several plots based on the real data
createRegionalPlots <- function (data, region) {
examplePlots <- list(ggplot() + ggtitle('Title (ggtext = plot 1)'),
ggplot() + ggtitle('Month (ggtext = plot 2)'),
ggplot() + ggtitle('Plot 1 (tile = 3)'),
ggplot() + ggtitle('Plot 2 (tile = 4)'),
ggplot() + ggtitle('Plot 3 (geom_bar = 5)'),
ggplot() + ggtitle('Plot 4 (geom_bar = 6)'),
ggplot() + ggtitle('Plot 5 (tile = 7)'),
ggplot() + ggtitle('Plot 6 (tile = 8)'))
}
# Found in https://stackoverflow.com/questions/43491685/
preparePage <- function(plots,layoutMatrix) {
# pdf(file = NULL) #invisible
par(mar=(c(5,5,5,5)))
plotsPerPage <- length(unique(na.omit(c(layoutMatrix))))
ml <- lapply(1:ceiling(length(plots)/plotsPerPage), function(page_IND){
ind <- (1 + ((page_IND - 1) * plotsPerPage )) : (page_IND * plotsPerPage)
grid.arrange(grobs = plots[ind], layout_matrix = layoutMatrix)
})
return(marrangeGrob(grobs=ml,nrow=1,ncol=1,top=NULL))
# dev.off() #invisible
}
# Here I'm running through all regions
regionalPlotList <- list()
for (region in names(criterias)) {
regionData <- criterias[[region]]
regionalPlots <- createRegionalPlots(data = regionData, region = region)
regionalPlotList <- do.call(c, list(regionalPlotList, regionalPlots))
}
# This leaves me with a list of 40 plots (5 regions x 8 plots)
allPlots <- preparePage(regionalPlotList, regionLayout)
ggsave("example.pdf",width = 297, height = 210, units = "mm", plot = allPlots)
As said, this works perfectly and leaves me (using the current data) with a five page report, one per every region and with the required layout.
I have now been asked to add additional per country plots at the end of the regional report and these pages should have a different layout (and different plots).
Overestimating myself (and my knowledge of r resp. ggplot) once again, I thought of this as an easy job (which it probably is for everyone else, but I'm stuck).
So, I've created a list of new criterias and a function, including a new layout:
createCountryPlots <- function(data, country) {
exampleCountryPlots <- list(ggplot() + ggtitle('Title (ggtext = plot 1)'),
ggplot() + ggtitle('Month (ggtext = plot 2)'),
ggplot() + ggtitle('Plot 1 (bar = 3)'),
ggplot() + ggtitle('Plot 2 (pie = 4)'),
ggplot() + ggtitle('Plot 3 (geom_bar = 5)'),
ggplot() + ggtitle('Plot 4 (geom_bar = 6)')
)
}
countryLayout = rbind(
c(1, 1, 1, 1, 1, 2),
c(3, 3, 3, 4, 4, 4),
c(3, 3, 3, 4, 4, 4),
c(3, 3, 3, 4, 4, 4),
c(5, 5, 5, 6, 6, 6),
c(5, 5, 5, 6, 6, 6),
c(5, 5, 5, 6, 6, 6)
)
# prepare the data per country
countryCriterias <- list()
countryCriterias[[ 'country_1' ]] <- data %>% filter(country == 'country 1')
countryCriterias[[ 'country_2' ]] <- data %>% filter(country == 'country 2')
# Running through all selected countries
countryPlotList <- list()
for (country in names(countryCriterias)) {
countryData <- countryCriterias[[country]]
countryPlots <- createCountryPlots(data = countryData, country = country)
countryPlotList <- do.call(c, list(countryPlotList, countryPlots))
}
countryPlots <- preparePage(countryPlotList, countryLayout)
# Just saving the country plots works perfectly again
ggsave("example.pdf",width = 297, height = 210, units = "mm", plot = countryPlots)
Saving this plots in a separate file works without any problems, but I'm currently stuck on how to combine these plots in one single PDF, respecting the different layouts the pages should have.
I've tried several possibilities (i.e. grid.arrange and arrangeGrob etc.), but I haven't been able to combine the plots into a single file.
Could anyone please enlighten me?
Edit:
Sorry, if I didn't make myself clear enough. This would be the result I should have at the end.

Answer 1

Thanks to the hint by #teunbrand to have a look at the patchwork package, I've found a solution to my problem.
It's in general almost the same as before, but instead of trying to arrange the plots first and then saving them, I "print" them directly to a pdf in the for-loop.
# defininig the layouts (simplified)
regionLayout <- "
AAAAAB
##CC##
DDEEFF
GGGHHH
GGGHHH"
countryLayout <- "
AAAAAB
CCCCDD
CCCCDD
EEEEFF
EEEEFF
"
# opening pdf
pdf('example5.pdf', pagecentre = FALSE, width = 29.7/2.54, height = 21/2.54)
par(mar = c(5, 5, 5, 5), oma = c(1, 1, 1, 1))
for (region in names(criterias)) {
regionData <- criterias[[region]]
regionalPlots <- createRegionalPlots(data = regionData, region = region)
# as regionalPlots is a list of plots, I'm using wrap_plots, which can take a dynamic
# number of plots
print(wrap_plots(regionalPlots, design = regionLayout))
}
# then the same for the country plots, with a different layout
countryPlotList <- list()
for (country in names(countryCriterias)) {
countryData <- countryCriterias[[country]]
countryPlots <- createCountryPlots(data = countryData, country = country)
print(wrap_plots(countryPlots, design = countryLayout))
}
dev.off()
And at the end I have my PDF with seperate layouts...
Thank you all for your help!!!
PS: Took me a while to find out why the PDF always was empty, before I realized that wrap_plot just arranges the plots but does not print them. As said, relatively new to R (did I mention that?)