Error in data == "grid": comparison (1) is possible only for atomic and list types

[DominiqueMakowski]

model <- lme4::lmer(Sepal.Length ~ Petal.Length + (1|Species), data=iris)

modelbased::estimate_link(modelbased::visualisation_matrix(model), model)
#> Error in data == "grid": comparison (1) is possible only for atomic and list types

^{Created on 2020-10-05 by the reprex package (v0.2.1)}

[strengejacke]

Why would you expect this to work? According to the docs (and code):

data

A data frame with model's predictors to estimate the response. If NULL, the model's data is used. If "grid", the model matrix is obtained (through visualisation_matrix).

So a model object should not work, or am I missing something?

[strengejacke]

library(lme4)
#> Loading required package: Matrix
library(modelbased)

model <- lme4::lmer(Sepal.Length ~ Petal.Length + (1|Species), data=iris)

estimate_link(visualisation_matrix(model))
#> Registered S3 method overwritten by 'broom.mixed':
#>   method      from 
#>   tidy.gamlss broom
#>    Petal.Length    Species Predicted    CI_low   CI_high
#> 1      1.000000     setosa  3.392931 0.8793172  5.906544
#> 2      1.655556     setosa  3.975373 1.3714259  6.579319
#> 3      2.311111     setosa  4.557815 1.8762909  7.239338
#> 4      2.966667     setosa  5.140257 2.4038720  7.876641
#> 5      3.622222     setosa  5.722699 2.9620915  8.483306
#> 6      4.277778     setosa  6.305141 3.5546777  9.055604
#> 7      4.933333     setosa  6.887583 4.1800520  9.595114
#> 8      5.588889     setosa  7.470025 4.8318077 10.108242
#> 9      6.244444     setosa  8.052467 5.5005394 10.604394
#> 10     6.900000     setosa  8.634909 6.1761700 11.093648
#> 11     1.000000 versicolor  3.392931 0.8793172  5.906544
#> 12     1.655556 versicolor  3.975373 1.3714259  6.579319
#> 13     2.311111 versicolor  4.557815 1.8762909  7.239338
#> 14     2.966667 versicolor  5.140257 2.4038720  7.876641
#> 15     3.622222 versicolor  5.722699 2.9620915  8.483306
#> 16     4.277778 versicolor  6.305141 3.5546777  9.055604
#> 17     4.933333 versicolor  6.887583 4.1800520  9.595114
#> 18     5.588889 versicolor  7.470025 4.8318077 10.108242
#> 19     6.244444 versicolor  8.052467 5.5005394 10.604394
#> 20     6.900000 versicolor  8.634909 6.1761700 11.093648
#> 21     1.000000  virginica  3.392931 0.8793172  5.906544
#> 22     1.655556  virginica  3.975373 1.3714259  6.579319
#> 23     2.311111  virginica  4.557815 1.8762909  7.239338
#> 24     2.966667  virginica  5.140257 2.4038720  7.876641
#> 25     3.622222  virginica  5.722699 2.9620915  8.483306
#> 26     4.277778  virginica  6.305141 3.5546777  9.055604
#> 27     4.933333  virginica  6.887583 4.1800520  9.595114
#> 28     5.588889  virginica  7.470025 4.8318077 10.108242
#> 29     6.244444  virginica  8.052467 5.5005394 10.604394
#> 30     6.900000  virginica  8.634909 6.1761700 11.093648

^{Created on 2020-10-05 by the reprex package (v0.3.0)}

[DominiqueMakowski]

Just in case you had a dataframe first and you attempted to use estimate_link in a chain:

data <- as.data.frame(modelbased::visualisation_matrix(lme4::lmer(Sepal.Length ~ Petal.Length + (1|Species), data=iris)))

data %>%
  modelbased::estimate_link(model)

[DominiqueMakowski]

so that essentially the two first arguments are interchangeable so that it doesn't error if you start the chain with data instead of model