Change some `tree_frogs` data variables to factor?

[juliasilge]

Would it be possible or worthwhile to update some of the variables in tree_frogs to be factors, specifically the ones that we use as outcomes for classification examples? This is necessary for computing metrics and would be nice if it were in the built-in data:

library(tidymodels)
library(stacks)

tree_frogs_class <- 
    tree_frogs %>%
    dplyr::select(-c(clutch, latency))

tree_frogs_class_split <- initial_split(tree_frogs_class)
tree_frogs_class_train <- training(tree_frogs_class_split)
tree_frogs_class_test  <- testing(tree_frogs_class_split)

log_st <-
    stacks() %>%
    add_candidates(log_res_nn) %>%
    add_candidates(log_res_rf) %>%
    blend_predictions() %>%
    fit_members()

augment(log_st, tree_frogs_class_test) %>%
    conf_mat(hatched, .pred_class)
#> Error in `yardstick_table()`:
#> ! `truth` must be a factor.
#> ℹ This is an internal error that was detected in the yardstick package.
#>   Please report it at <https://github.com/tidymodels/yardstick/issues> with a reprex (<https://https://tidyverse.org/help/>) and the full backtrace.

#> Backtrace:
#>     ▆
#>  1. ├─augment(log_st, tree_frogs_class_test) %>% ...
#>  2. ├─yardstick::conf_mat(., hatched, .pred_class)
#>  3. └─yardstick:::conf_mat.data.frame(., hatched, .pred_class)
#>  4.   └─yardstick:::yardstick_table(truth = truth, estimate = estimate, case_weights = case_weights)
#>  5.     └─rlang::abort("`truth` must be a factor.", .internal = TRUE)

^{Created on 2023-01-25 with reprex v2.0.2}

I suspect you did this manually when creating the example/test objects like log_res_nn? For example, when I see this in hatched:

library(tidymodels)
library(stacks)
log_res_nn %>%
    collect_predictions()
#> # A tibble: 909 × 7
#>    id    .pred_no .pred_yes  .row .pred_class hatched .config             
#>    <chr>    <dbl>     <dbl> <int> <fct>       <fct>   <chr>               
#>  1 Fold1    0.469     0.531     1 yes         yes     Preprocessor1_Model1
#>  2 Fold1    0.725     0.275     5 no          no      Preprocessor1_Model1
#>  3 Fold1    0.572     0.428    16 no          no      Preprocessor1_Model1
#>  4 Fold1    0.721     0.279    17 no          no      Preprocessor1_Model1
#>  5 Fold1    0.503     0.497    28 no          no      Preprocessor1_Model1
#>  6 Fold1    0.689     0.311    29 no          no      Preprocessor1_Model1
#>  7 Fold1    0.464     0.536    31 yes         no      Preprocessor1_Model1
#>  8 Fold1    0.270     0.730    32 yes         yes     Preprocessor1_Model1
#>  9 Fold1    0.270     0.730    35 yes         yes     Preprocessor1_Model1
#> 10 Fold1    0.495     0.505    38 yes         no      Preprocessor1_Model1
#> # … with 899 more rows

^{Created on 2023-01-25 with reprex v2.0.2}

[simonpcouch]

cc @topepo, who has brought this up in the past as well.

Regarding the example objects, that coercion to factor happens somewhere (tune? hardhat?) automatically outside of stacks.

I'm on board for converting the current character variables to factor. :)

[github-actions]

This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.