Can't use a y variable in aes_string

Question

I am currently working with COVID-19 data from Peru and I want to use a Rmisc::multiplot to show the development of positive cases in each region of the country. So I am trying to code a loop to generate 25 plots. Only as a matter of example I am only using 4 variables:

     Fecha    Lima     La Libertad  Madre de Dios
1 2020-04-24   10           2             1
2 2020-04-25   15           4             3
3 2020-04-26   20           8             3

I generated a vector with the names of the regions:

nombre_regiones <- c("Lima", "La Libertad", "Madre de Dios")

And I created an empty list to store the 25 plots within:

regiones <- list()

Then I used this for loop to generate and store each plot in the list "regiones":

for (w in seq_along(nombre_regiones)) { 
  nombre <- paste("r", w, sep = "")
  assign(nombre, ggplot(data = df, aes_string(x = "Fecha", y = nombre_regiones[w])) + geom_line() + geom_point() + scale_x_date(date_labels = "%d, %m", date_breaks  ="1 day") + geom_text(aes_string(label = nombre_regiones[w])))
  regiones[[w]] <- nombre
}

The loop created r1 and stored the plot within but when w = 2, that means nombre_regiones[w] = "La Libertad" I got the next warning:

Error in parse(text = x) : <text>:1:4: unexpected symbol
1: La Libertad
       ^

Because of that I can't create the second plot and the same thing happens with the third region "Madre de Dios". I tried different things and look up for similar cases but I didn't have luck. Also regiones[[w]] <- nombre isn't working but that's something I'll look up later. Thanks in advance.

The dataframe with the variables "Fecha", "Lima", "La Libertad" and "Madre de Dios" is store with the object name "df". — Ernesto Riveros Barrientos, Apr 27 '20 at 01:16
Please check my solution below. It is working with the changes. — akrun, Apr 27 '20 at 01:22
@akruni got this in the console `Error: Invalid input: date_trans works with objects of class Date only` when I try to call r1, r2 or r3. — Ernesto Riveros Barrientos, Apr 27 '20 at 03:19
you probbaly need to run `df$Fecha <- as.Date(df$Fecha)` first. — Ronak Shah, Apr 27 '20 at 03:58
@ErnestoRiverosBarrientos if you check my data, the `Fecha` is `Date` class. — akrun, Apr 27 '20 at 17:35
The problem is I uses this `df$Fecha <- as.Date(df$Fecha, origin = "1899-12-30")` to prepare my data and when I erase the `origin` parameter I got `Error in as.Date.numeric(df$Fecha) : 'origin' must be supplied`. So I think the problem is there. — Ernesto Riveros Barrientos, Apr 27 '20 at 19:00

score 0 · Accepted Answer · edited Jun 20 '20 at 09:12

It is better to initiate the list with full length

regiones <- vector('list', length(nombre_regiones))

Also, the aes_string is now deprecated. Instead, can be converted to symbol and evaluated (!!)

library(ggplot2)
for (w in seq_along(nombre_regiones)) { 
      nombre <- paste0("r", w)
      assign(nombre, ggplot(data = df,
          aes(x = Fecha, y = !! rlang::sym(nombre_regiones[w]))) +   
              geom_line() + 
              geom_point() + 
              scale_x_date(date_labels = "%d, %m", date_breaks  ="1 day") +
              geom_text(aes(label = !! rlang::sym(nombre_regiones[w]))))
      regiones[[w]] <- nombre
 }

-output r2

-output r3

data

df <- structure(list(Fecha = structure(c(18376, 18377, 18378), class = "Date"), 
    Lima = c(10L, 15L, 20L), `La Libertad` = c(2L, 4L, 8L), `Madre de Dios` = c(1L, 
    3L, 3L)), row.names = c("1", "2", "3"), class = "data.frame")

I forgot to change the `paste` function. It ran as I wanted, thank you very much. — Ernesto Riveros Barrientos, Apr 27 '20 at 21:15

Can't use a y variable in aes_string

1 Answers1

data