make multiple separate stacked barplots from one data frame - r

Question

I would like to create multiple grouped and stalked barplots with several data frames and be able to export the plots i a single pdf file.
I have several data frames with the same format but varying values. For each data frame I would like to create multiple stalked and grouped bar plots. Ideally the bar plots of the same group from the data frames should be placed next to each other and share the same Y-axis length (in order to easily visually compare the data frames).
Her an example of what ma data looks like:
data1 <- data.frame(group=c('A','A','A','A','B','B','B','B','C','C','C','C'),
Year=c('2012','2013','214','2015','2012','2013','214','2015','2012','2013','214','2015'),
Fruit=c(5,3,6,3,5,4,2,2,3,4,6,2),
Vegetables=c(3,6,1,4,8,9,43,2,1,5,0,1),
Rice=c(20,23,53,12,45,5,23,12,32,41,54,32))
data2 <- data.frame(group=c('A','A','A','A','B','B','B','B','C','C','C','C'),
Year=c('2012','2013','214','2015','2012','2013','214','2015','2012','2013','214','2015'),
Fruit=c(2,4,5,2,3,9,4,7,5,7,4,7),
Vegetables=c(9,7,8,7,4,3,0,0,2,3,5,6),
Rice=c(23,12,32,41,54,32,20,23,53,12,45,5))
data1 <- pivot_longer(data1, cols = 3:5, names_to = 'Type', values_to = 'value')
data2 <- pivot_longer(data1, cols = 3:5, names_to = 'Type', values_to = 'value')
I started by formating the tables like this:
data1 <- pivot_longer(data1, cols = 3:5, names_to = 'Type', values_to = 'value')
data2 <- pivot_longer(data1, cols = 3:5, names_to = 'Type', values_to = 'value')
My attempts to use ggplot to create the desired PDF have so far failed. I took several different attempts but could not get near to the desired PDF. I found instructions on how to create several plots for one data frame, or grouped plots or stalked plots, but never the combination of all three.
If possible the PDF I would like to get for this example should look like this:
In total 6 plots: left 3 plots data1, right 3 plots data2; Group A row1, Group B row2, Group C row3 (if possible same y axis length in one row/Group)
All bar plots: x-axis= years, y-axis= value / 1 stalked bar per year with colors matching Type (Fruit, Vegetable, Rice)
Group name per row
data source(data1, data2) per column
legend with Types (Fruit, Vegetable, Rice)
Q1. Is something like this possible or would one have to create two PDFs (for each data.table, here: data1 and data2).
Q2. Is it possible to format the code in a way to automatically adjust the amount of plots needed according to the data frames and adjust the PDF file size automatically and create a new page if necessary? (In reality i have 5 data frames and 13 Groups, this may however change with time)
I know this is quite a difficult code to write. I have spent two working days on this already though, which is why I am now asking for help here. I will try again tomorrow and post any possible progress here.
Thank you very much for any suggestions

Answer 1

This code should produce the desired plot (or at least something really close).The two critical steps include: 1) joining all the dataframes into a single one, using bind_rows and 2) using facet_grid to set define the layout panels according to two variables (group and id).
library(tidyverse)
# Combine the data
# id column contains the number of the dataframe from which the data comes from
df <- bind_rows(data1, data2, .id = "id")
df %>%
# Change to long format, add 1 to the columns number, as we now added id column
pivot_longer(cols = 4:6,
names_to = 'Type',
values_to = 'value') %>%
# Transform value to x / 1
mutate_at(vars(value), function(x) x / 1) %>%
# Do plot
ggplot(aes(x = Year,
y = value,
fill = Type)) +
# columns
geom_col()+
# Facets by two factors, groups and data source (id)
facet_grid(group ~ id)
# Save plot to pdf
ggsave("my_plot.pdf",
device = "pdf",
height = 15,
width = 20,
units = "cm",
dpi = 300)