CODEWITHSUNDEEP
rgraphsjplot
Ruben
Ruben

Reputation: 1

Fit several predictions plots of sjPlot in one window

I've come across a problem when using sjPlot using the "pred" function. With the model I currently have, three predictors are used in the model.

NQuart3.1 <- 
  glmmTMB(Pos_count ~ Predsumscale + Buildscale + 
          (1|Territory),
          offset = log(Ndays),
          dispformula = ~Roadscale + Buildscale, 
          data = NightQuart3, 
          family = nbinom1(link = "log")
          )

when I use sjplot on this, it will provide me with 2 separate graphs. One for Buildscale and one for Predsumscale, separated.

plot_model(NQuart3.1, type = "pred",
           title = "Wolf space use night estimates April - June",
           axis.title  = "Wolf GPS locations")

$Predsumscale
(http://127.0.0.1:26952/graphics/ede13814-0fa4-4ba3-8c20-ae47c68dcabc.png)
$Buildscale 
(http://127.0.0.1:26952/graphics/fe656c83-5bb0-4692-ad35-dd23a7ce3f0a.png)

I was wondering if you can adjust the code so that it would combine both graphs in one plot window?

Not combined but just side-by-side.

This would make it a lot easier to compare models with multiple parameters.

Thanks in advance

Upvotes: 0

Views: 939

Answers (1)

Ben Bolker
Ben Bolker

Reputation: 226662

I don't know of a way to tell plot_model() explicitly "please put all of these in the same plot window", but there are many easy ways to take the plot object and render the components within a single window with different layouts.

Setup:

library(glmmTMB)
m1 <- glmmTMB(mpg ~ disp + hp, mtcars)
library(sjPlot)
pp <- plot_model(m1, type = "pred")

sjPlot has a plot_grid(), but it doesn't seem very flexible.

sjPlot::plot_grid(pp)

or

gridExtra::grid.arrange(grobs=pp)

cowplot has a plot_grid() with many more options

cowplot::plot_grid(plotlist = pp)

patchwork is probably the most flexible.

library(patchwork)
pp[[1]] + pp[[2]]

Upvotes: 0

Related Questions