Fix "Error: Discrete value supplied to continuous scale" in R

June 5, 2024

I have a data frame that was returned after running a model that contains production values, the mean profit, and upper/lower boundaries for 80% and 90% confidence intervals. From there, I’m trying to create an overlapping line plot with the mean and ribbon plot for the confidence intervals. When running the code below, I end up getting the message "Error: Discrete value supplied to continuous scale"

#code from assignment 3
set.seed(114)
n <- 100000
probs <- c(.3,.45,.2,.05)
ints <- c(2000,5001,10001,14001,15000)

qbp <- function(p,probs,ints){
  ifelse(p<probs[1],
         qunif(p/probs[1],ints[1],ints[2]),qbp(p-probs[1],probs[-1],ints[-1]))
}

rbp <- function(n,probs,ints){
  p <- runif(n)
  sapply(p,qbp,probs=probs,ints=ints)
}

boat_df <- data.frame(
  demand = rbp(n,probs,ints) %>% round(),
  cost_fix = rnorm(n,300000000,60000000) %>% round(2),
  cost_var = rtriang(n,77000,100000,90000) %>% round(2)
)


#model updated to return only profit
boat_sim <- function(prod){

retail_full <- 150000
retail_disc <- 75000

boat_df <- boat_df %>%
  mutate(
    units_full = ifelse(prod > demand, demand, prod),
    units_disc = ifelse(prod > demand, prod-demand, 0),
    revenue = units_full*retail_full + units_disc*retail_disc,
    cost = cost_fix + prod*cost_var,
    profit = revenue - cost
  )
return(boat_df$profit)
}

#generate summary results
prod_1a <- seq(2000,15000,1000)

sum_stats <- function(prod_1a){
  x <- boat_sim(prod_1a)
  return(
    c(prod_1a=prod_1a,
      mean=mean(x),
      quantile(x,c(.005,.100,.900,.995))))
}

summary_1a <-
  sapply(prod_1a,sum_stats) %>%
  t() %>%
  as.data.frame()

#line & ribbon plot
ggplot(summary_1a)+
  geom_line(aes(x=prod_1a,
                y=mean))+
  geom_ribbon(aes(x=prod_1a,
                  ymin='0.5%',
                  ymax='99.5%'),
              alpha=.3)+
  geom_ribbon(aes(x=prod_1a,
                  ymin='10%',
                  ymax='90%'),
              alpha=.3)

From searching the issue, it seems like everyone is saying that I need to ensure that all of the variables are numeric, but when I run str(summary_1a) it shows that they all are numeric so I’m not sure where I’m going wrong with my code.

'data.frame':   14 obs. of  6 variables:
 $ prod_1a: num  2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 ...
 $ mean   : num  -1.78e+08 -1.21e+08 -7.12e+07 -2.91e+07 5.90e+06 ...
 $ 0.5%   : num  -3.35e+08 -2.86e+08 -2.80e+08 -2.96e+08 -3.13e+08 ...
 $ 10%    : num  -2.56e+08 -2.02e+08 -1.65e+08 -1.53e+08 -1.64e+08 ...
 $ 90%    : num  -1.00e+08 -4.07e+07 1.79e+07 7.58e+07 1.33e+08 ...
 $ 99.5%  : num  -21006018 40849682 102371443 163326909 225063557 ...

>Solution :

Replace the ticks (‘ ‘) with backticks (“). You can also simplify your code by putting common aesthetics in the ggplot call.

ggplot(summary_1a, aes(x=prod_1a)) +
  geom_line(aes(y=mean)) +
  geom_ribbon(aes(ymin=`0.5%`,
                  ymax=`99.5%`),
              alpha=.3)+
  geom_ribbon(aes(ymin=`10%`,
                  ymax=`90%`),
              alpha=.3)