Issue plotting lme4 random effects from model_parameters

[DanSimonet]

Amazing suite of packages! I was hoping to pivot my entire mixed modeling workflow to easystats but came across an issue in how model_parameters extracts parameter labels (I believe from broom.mixed) which prevents plotting of group-specific random effects. I am expecting output similar to sjPlot and use this as comparison.

Model

library(lme4)
s_mod <- lmer(Reaction ~ Days + (Days|Subject), data = sleepstudy)

Expectation

# Sjplot output
sjPlot::plot_model(s_mod, show.values = T, type = "re", value.offset = 0.5)

Model Parameter + See Result

# Parameters -> see output
parameters::model_parameters(s_mod, group_level = TRUE) |> plot()

Source
The problem I believe is broom.mixed::tidy returns term and level as different columns. model_parameters as a result has a single value for the grouping variable in the Parameter column as opposed to the group-specific identifiers which show up in Level.

[strengejacke]

I think it's because we just haven't a proper plot() method for random effects when group_level = TRUE. Let me check.

[strengejacke]

Rendering is not perfect here, but these are the options:

library(lme4)
#> Loading required package: Matrix
s_mod <- lmer(Reaction ~ Days + (Days | Subject), data = sleepstudy)

parameters::model_parameters(s_mod, group_level = TRUE) |>
  plot()

parameters::model_parameters(s_mod, group_level = TRUE) |>
  plot(show_labels = TRUE)

parameters::model_parameters(s_mod, group_level = TRUE) |>
  plot(show_labels = TRUE, size_text = 5)

parameters::model_parameters(s_mod, group_level = TRUE) |>
  plot(sort = "ascending", show_labels = TRUE)

parameters::model_parameters(s_mod, group_level = TRUE) |>
  plot(sort = "ascending", show_labels = TRUE, n_columns = 2)

parameters::model_parameters(s_mod, group_level = TRUE) |>
  plot(sort = "ascending", show_labels = TRUE, show_intercept = FALSE)

^{Created on 2024-09-27 with reprex v2.1.1}