Builds an exposure-response plot for a logistic regression model

Usage

lr_plot(data, exposure, response, stratify_by = NULL)

lr_plot_style(object, labels)

lr_plot_show_model(
  object,
  keep_strata = NULL,
  style = "ribbonline",
  conf_level = 0.95
)

lr_plot_show_quantiles(
  object,
  keep_strata = NULL,
  style = "errorbar",
  bins = 4,
  conf_level = 0.95
)

lr_plot_show_datastrip(
  object,
  keep_strata = NULL,
  style = "jitter",
  panel = "both"
)

lr_plot_show_groups(
  object,
  group_by,
  style = "boxplot",
  bins = NULL,
  keep_strata = NULL
)

lr_plot_build(object)

Arguments

data: Observed data
exposure: Exposure variable (one variable, unquoted)
response: Response variable (one variable, unquoted)
stratify_by: Stratification variable used for color and fill (one variable, unquoted)
object: Partially constructed plot (has S3 class erlr_plot)
labels: Named list of labels
keep_strata: Logical, indicating whether this component should keep the color stratification
style: Character string used to specify the partial builder for this component
conf_level: Confidence level
bins: Number of exposure bins (not counting placebo)
panel: Character string: "upper", "lower", or "both" (the default)
group_by: Grouping variables to define groups for distribution plots (a tidyselection of variables)

Value

Plot object of class erlr_plot

Examples

lr_data |> 
  lr_plot(aucss, ae1) |> 
  lr_plot_show_model() |> 
  lr_plot_show_quantiles() |> 
  lr_plot_show_groups(aucss) |> 
  plot()

 
plt <- lr_data |> 
  lr_plot(aucss, ae2, stratify_by = sex) |> 
  lr_plot_show_model(keep_strata = FALSE) |> 
  lr_plot_show_quantiles() |> 
  lr_plot_show_datastrip() |> 
  lr_plot_show_groups(group_by = c(aucss, treatment), keep_strata = FALSE)

print(plt)
#> <erlr_plot>
#>   plot variables:
#>     - exposure:        aucss
#>     - response:        ae2
#>     - stratification:  sex
#>   plot components:
#>     - model:           ae2 ~ aucss
#>     - quantile:        4 bins
#>     - strip:           jitter both
#>     - group:           .aucss_quantile, treatment
#>   plots built: <none>
#>   output built: no
plot(plt)