When I run this order using the forecast package, the following Facet Chart is currently produced.

library(fabletools)
library(forecast)

Test_plt<-fdeaths %>% 
  forecast()

checkresiduals(Test_plt)

Although Forecast works with ggplot2, what it produces cannot be called again as an object. For example, if I click on 'Test_plt', I will only see the test results conducted in the model but not the graph itself.

plot1<-checkresiduals(Test_plt)

 > plot1

    Ljung-Box test

data:  Residuals from ETS(M,N,M)
Q* = 11.592, df = 14, p-value = 0.639

Can anyone replicate this graph as a ggplot2 object so that it can be called later?

When I run this order using the forecast package, the following Facet Chart is currently produced.

library(fabletools)
library(forecast)

Test_plt<-fdeaths %>% 
  forecast()

checkresiduals(Test_plt)

Although Forecast works with ggplot2, what it produces cannot be called again as an object. For example, if I click on 'Test_plt', I will only see the test results conducted in the model but not the graph itself.

plot1<-checkresiduals(Test_plt)

 > plot1

    Ljung-Box test

data:  Residuals from ETS(M,N,M)
Q* = 11.592, df = 14, p-value = 0.639

Can anyone replicate this graph as a ggplot2 object so that it can be called later?

• r
• ggplot2
• forecast

1 Answer 1

I don't see any easy option to return all plots and store them in a variable or as a list for further manipulation. But after a look at the source code of ggtsdisplay which is called under the hood by ckeckresiduals you can re-create the three subplots like so:

library(fabletools)
library(forecast)
library(ggplot2)

Test_plt <- fdeaths |>
  forecast()

checkresiduals(Test_plt)

tsplot <- autoplot(Test_plt$residuals) +
  geom_point(size = 0.5) +
  labs(y = NULL)

acfplot <- ggAcf(Test_plt$residuals, lag.max = 24) +
  labs(title = NULL)

histplot <- gghistogram(Test_plt$residuals, add.normal = TRUE, add.rug = TRUE) + 
  labs(x = "residuals")

library(patchwork)

tsplot / (acfplot + histplot)