Multiple individual graphs from a unique dataframe - r

Question

I know that some subjects are about similar questions, but even using those I was not able to resolve the issue on my own. Thus, I am sorry if this subject appears as a duplicate but I am a bit stuck.
I have to draw nearly 40 graphs representing body temperature variations accross 24hours (a graph per individual of the study). To do that, I tried to write a loop using dplyrand ggplot2 packages. You may find bellow an exemple of my data. There are numerous missing values but I don't think it represents an issue regarding the current question.
structure(list(heures = structure(1:13, .Label = c("01:00:00",
"03:00:00", "05:00:00", "07:00:00", "08:00:00", "10:00:00", "12:00:00",
"13:30:00", "15:00:00", "17:00:00", "19:00:00", "21:00:00", "23:00:00"
), class = "factor"), x1= c(36.55, 36.5, 36.44444444,
36.6, 36.86666667, 37.26, 37, NA, NA, 37.3, 37.1, 37, 35.6),
x2 = c(NA, 34.5, 35.4, 36.1, NA, NA, NA, NA, NA,
NA, NA, NA, NA), x3 = c(36.9, 36.4, NA, NA, 36.9,
NA, NA, NA, NA, 37.5, 37.5, 36.9, 37.1), x4 = c(36,
35.8, NA, NA, NA, 37.4, 36.7, 36.3, NA, 37.5, 37, NA, NA)), class = "data.frame", row.names = c(NA,
-13L))
So far, I have written the following code with "indiv" being a dataframe containing the above presented data.
names <- c(colnames(indiv))
graph <- list()
test <- function(df, names) {
for (i in 1:length(df)) {
name <- names[i]
stock <- df %>%
filter(heures, !!name)
graph[[i]] <- ggplot(data=stock, aes(x=heures, y=stock[,2])) +
geom_point() +
labs(x="Hours (HH:MM:SS)",
y="Temperature",
title=colnames(stock[2]))
}
return(graph)
}
It returns an error that seems to indicate the filter function does not work properly:
Warning messages:
1: In Ops.factor(~heures, ~"x1") :
‘&’ not meaningful for factors
I can't figure out what I'm doing wrong in this. I also tried a code without the dplyr part present in the current loop, but it didn't gave me the wanted output neither.
Thank you in advance for your advises.

Answer 1

I have came out with this idea : tidy a little the dataset to make it easier to use with ggplot and then split it and store the splitted dataframe in a list. Then I use lapply to avoid using a loop along with a custom function to create plots.
This is not a very fast way if you have a lot of data but I use this trick a lot with small datasets.
This code creates a plot for each individual (not facets).
library(tidyverse) # all functions of these packages are not necessary here
df = structure(list(heures = structure(1:13, .Label = c("01:00:00",
"03:00:00", "05:00:00", "07:00:00", "08:00:00", "10:00:00", "12:00:00",
"13:30:00", "15:00:00", "17:00:00", "19:00:00", "21:00:00", "23:00:00"
), class = "factor"), x1= c(36.55, 36.5, 36.44444444,
36.6, 36.86666667, 37.26, 37, NA, NA, 37.3, 37.1, 37, 35.6),
x2 = c(NA, 34.5, 35.4, 36.1, NA, NA, NA, NA, NA,
NA, NA, NA, NA), x3 = c(36.9, 36.4, NA, NA, 36.9,
NA, NA, NA, NA, 37.5, 37.5, 36.9, 37.1), x4 = c(36,
35.8, NA, NA, NA, 37.4, 36.7, 36.3, NA, 37.5, 37, NA, NA)), class = "data.frame", row.names = c(NA,
-13L))
# tidy your data, good practice makes it easier to plot things with ggplot
df = df %>% pivot_longer(2:ncol(df), names_to = "individual", values_to = "temperature")
# I would do it this way:
df_list = split(df, df$individual)
plot_fun = function(df) {
title = unique(df$individual)
ggplot(df, aes(x=heures, y=temperature))+
geom_point() +
labs(title = title)
#### add here things to save your plots, store them somewhere, etc
}
lapply(df_list, FUN = plot_fun)

Answer 2

Using toy data as your data frame is incomplete:
df <- tibble(
X=rep(1:10, times=2),
Y=c(1:10, seq(10, 1, -1)),
Name=rep(c("Patient 1", "Patient 2"), each=10)
)
df %>% ggplot() +
geom_line(aes(x=X, y=Y)) +
facet_grid(rows=vars(Name))
Giving