"Error: Invalid input: time_trans works with objects of class POSIXct only" with geom_rect and date/POSIXct types

Question

I have a dataset as follows:

# A tibble: 30 x 4
# Groups:   Group, Week_End_Date [30]
   Group Week_End_Date       time_period             Outcome
   <lgl>   <dttm>              <fct>                   <dbl>
 1 FALSE   2019-09-22 00:00:00 Baseline                5241.
 2 FALSE   2019-09-29 00:00:00 Baseline                5089.
 3 FALSE   2019-10-06 00:00:00 Baseline                4991.
 4 FALSE   2019-10-13 00:00:00 Baseline                4920.
 5 FALSE   2019-10-20 00:00:00 Baseline                4896.
 6 FALSE   2019-10-27 00:00:00 Baseline                4921.
 7 FALSE   2019-11-03 00:00:00 Baseline                4999.
 8 FALSE   2019-11-10 00:00:00 Activation              5003.
 9 FALSE   2019-11-17 00:00:00 Activation              4995.
10 FALSE   2019-11-24 00:00:00 Activation              5098.
# ... with 20 more rows

I can plot a graph e.g.

    ggplot(dataset, aes(y=Outcome, x=Week_End_Date, color=Group, fill=Group)) +
      geom_point() +
      geom_line() +
      theme_bw() +
labs(y="Outcome", x="Date", color="Group", fill="Group")

I want to shade this based on the variable "time_period" (Week_End_Date is nested within this)

e.g.

ggplot(dataset, aes(y=Outcome, x=Week_End_Date, color=Group, fill=Group)) +
  geom_point() +
  geom_line() +
  theme_bw()
  geom_line() +
  theme_bw() +
 geom_rect(aes(fill=time_period, xmax=Inf, xmin=-Inf, ymax=Inf, ymin=-Inf)) 

however this results in the error message

Error: Invalid input: time_trans works with objects of class POSIXct only

I also tried drawing the two rectangles separately (I will add the alpha argument later once they plot successfully):

    ggplot(dataset, aes(y=Outcome, x=Week_End_Date, color=Group, fill=Group)) +
      geom_point() +
      geom_line() +
      theme_bw()
      geom_line() +
      theme_bw() +
geom_rect(inherit.aes=FALSE, aes(xmin=dmy("06Nov2019"), xmax=dmy("06Jan2020"), ymax=Inf, ymin=-Inf), fill="red") +
  geom_rect(inherit.aes=FALSE, aes(xmin=dmy("01Jan2019"), xmax=dmy("06Nov2019"), ymax=Inf, ymin=-Inf), fill="blue") 

but this results in the same error.

Can anyone advise me on what I'm doing wrong? I feel like this should be far more straightforward than it seems! The x variable is a datetime

Answer 1

I try to simulate something that looks like your dataset, but as @camille said, try to provide a dataset to go with your code? This makes things way faster.

First the data:

DATES = seq(as.POSIXlt("2019-01-01 00:00:00", tz="UTC"), 
    as.POSIXlt("2020-01-31 00:00:00", tz="UTC"), length.out=20)

dataset = data.frame(
  Outcome = c(rpois(20,1000),rpois(20,2000)),
  Week_End_Date = rep(DATES,2),
  Group = rep(c(FALSE,TRUE),each=20)
)
dataset$time_period = ifelse(
dataset$Week_End_Date < as.POSIXlt("2019-11-06 00:00:00", tz="UTC"),
"Baseline",
"Activation"
)

What you need is another data frame that specifies the start and end of your two time periods:

FRAME = dataset %>% group_by(time_period) %>% 
summarize(xmin=min(Week_End_Date),xmax=max(Week_End_Date))

Then similar to what you have done, we can use this in geom_rect, except that now, it comes with a legend:

ggplot(dataset, aes(y=Outcome, x=Week_End_Date, color=Group)) +
  geom_point() +
  geom_line() +
  theme_bw() +
  labs(y="Outcome", x="Date", color="Group", fill="Group")+
  geom_rect(data=FRAME,inherit.aes=FALSE,
  aes(xmin=xmin,xmax=xmax,ymax=+Inf,ymin=-Inf,fill=time_period),alpha=0.1)+
  scale_fill_manual(values=c("red","blue"))