R – Replicate/Automate scatterplots for each string group in column - r

Question

I have a dataframe (170 observation and 3 columns) like so:
site_names <- c("WS1", "WS1", "WS2", "WS2", "WS3", "WS3")
x <- c(.15, .20, .17, .16, .20, .22)
y <- c(.026, .031, .045, .087, .09, .033)
df <- data.frame(site_names, x, y)
and I have a simple ggplot formula I can run for each site_name where I plot x and y accordingly:
df %>%
filter(site_names == "") %>%
ggplot(aes(x = x, y = y)) +
geom_point() +
labs(x = "Discharge", y = "Stage")
What I would like is to automate these plots and generate 26 separate plots by site_name and save them all to a PDF
Here is what I have tried:
scatter_expl = function(x, y) {
ggplot(df, aes(x = x, y = y) ) +
geom_point() +
theme_bw() +
labs(x = "Discharge", y = "Stage")
}
plots = map(df, ~scatter_expl())
plots
But this runs through the whole dataframe, returns a list and produces all scatters on the same plot: https://i.stack.imgur.com/gVBhf.png! How do I group by site and return individual graphs?

Answer 1

We may need to group_by
library(dplyr)
library(ggplot2)
library(gridExtra)
out <- df %>%
group_by(site_names) %>%
summarise(plot = list(ggplot(cur_data(), aes(x = x, y = y)) +
geom_point() +
labs(x = "Discharge", y = "Stage")+
ggtitle(sprintf('Plot for %s', cur_group()$site_names))))
-output
> out
# A tibble: 3 × 2
site_names plot
<chr> <list>
1 WS1 <gg>
2 WS2 <gg>
3 WS3 <gg>
-save the output as pdf
ggsave(file.path(getwd(), "plot.pdf"),
marrangeGrob(out$plot, nrow = 1, ncol = 1), device = "pdf")
data
df <- structure(list(site_names = c("WS1", "WS1", "WS2", "WS2", "WS3",
"WS3"), x = c(0.15, 0.2, 0.17, 0.16, 0.2, 0.22), y = c(0.026,
0.031, 0.045, 0.087, 0.09, 0.033)), class = "data.frame", row.names = c(NA,
-6L))

Answer 2

Updated with the help of dear #akrun:
library(tidyverse)
library(purrr)
library(gridExtra)
# create list of df by site_names
dflist <- df %>%
group_split(site_names)
# create a function to plot
scatter_fun = function(dat) {
ggplot(dat, aes(x = x, y = y)) +
geom_point() +
theme_bw() +
labs(x = "Discharge", y = "Stage")
}
# iterate through each list element and apply function (many thanks to #akrun for his help)
p <- map(dflist, ~scatter_fun(.x))
# save all in one pdf
ggsave(
filename = "plots.pdf",
plot = marrangeGrob(p, nrow=2, ncol=2),
width = 7.5, height = 4.5
)