How to change the legend title of my ggplots based on outcomes? - r

Question

I wonder if there is an efficent way to change the legend title of my ggplots based on outcomes.
Example
I have a function which helps me to filter a data base by type and by county.
df without filter
county | date | value | type
-----------------------------------
Alameda 2020-01-01 6 positive
Alameda 2020-01-02 2 negative
Alameda 2020-01-03 1 positive
LA 2020-01-04 4 positive
LA 2020-01-03 1 positive
** Function **
function_forggplot <- function(data = df,
select_county = "Alameda",
type_order = unique(df$type)) {
#Filter data base
df_outcome <- df[df$county %in% select_county,]
df_outcome <- df_outcome[df_outcome$type %in% type_order,]
gg_outcome <- ggplot(
data = df_outcome,
aes(x = date,
y = value,
color = type
)) +
geom_line(size = .5)
I want change the legend title of my ggplot2 based on the outcome for example if the user select county= LA and type = positive. I want a title in my ggplot like "Results positives for LA".
With if else conditionals works but I have more than 100 cases so I think this not a good option.
Expect outcome
function_forggplot(county="Alameda", type = "negatives")
A ggplot object with this title "Results negatives for Alameda"
function_forggplot(county="Fresno", type = "postives")
A ggplot object with this title "Results pisitives for Fresno"
Thanks

Answer 1

You can adapt the code you showed in the function directly like this. Also you can use paste0() for the title (no need of other packages):
library(ggplot2)
#Function
funplot <- function(df, select_county, type_order) {
#Filter data base
df_outcome <- df[df$county == select_county,]
df_outcome <- df_outcome[df_outcome$type == type_order,]
#Plot
ggplot(df_outcome,aes(x = date, y = value)) +
geom_line() +
ggtitle(paste0("Results ",select_county," for ",type_order))
}
#Apply
funplot(mydf, 'Alameda', 'positive')
Output:
Some data used:
#Data
mydf <- structure(list(county = c("Alameda", "Alameda", "Alameda", "LA",
"LA"), date = structure(c(18262, 18263, 18264, 18265, 18264), class = "Date"),
value = c(6L, 2L, 1L, 4L, 1L), type = c("positive", "negative",
"positive", "positive", "positive")), row.names = c(NA, -5L
), class = "data.frame")

Answer 2

We can create the function with glue as it is very flexible to get objects specified within {}. Of course, we can use paste or sprintf as well. Anyway, ggplot is an external package. So using, another package from the tidyverse, would make this more tidier.
library(ggplot2)
library(dplyr)
f1 <- function(dat, county_nm, type_nm) {
dat %>%
filter(county == county_nm, type == type_nm) %>%
ggplot(aes(x = date, y = value)) +
geom_line() +
ggtitle(glue::glue("Results {type_nm} for {county_nm}"))
}
then, we call as
f1(df, 'LA', 'positive')
-output
Or without any packages (other than ggplot2, dplyr)
f2 <- function(dat, county_nm, type_nm) {
dat %>%
filter(county == county_nm, type == type_nm) %>%
ggplot(aes(x = date, y = value)) +
geom_line() +
ggtitle(sprintf("Results %s for %s", type_nm, county_nm))
}
data
df <- structure(list(county = c("Alameda", "Alameda", "Alameda", "LA",
"LA"), date = structure(c(18262, 18263, 18264, 18265, 18264), class = "Date"),
value = c(6L, 2L, 1L, 4L, 1L), type = c("positive", "negative",
"positive", "positive", "positive")), row.names = c(NA, -5L
), class = "data.frame")