Plot a Decision Tree from Margot Causal-Forest Results — margot_plot_decision

Creates a visualisation of a policy tree showing the decision rules and assigned treatments from causal forest results.

Usage

margot_plot_decision_tree(
  result_object,
  model_name = NULL,
  max_depth = 2L,
  original_df = NULL,
  x_padding = 0.12,
  y_padding = 0.25,
  border_size = 0.5,
  text_size = 4,
  edge_label_offset = 0.025,
  span_ratio = 0.4,
  non_leaf_fill = "lightyellow",
  title = NULL,
  plot_margin = grid::unit(c(1, 1, 1, 1), "cm"),
  remove_tx_prefix = TRUE,
  remove_z_suffix = TRUE,
  use_title_case = TRUE,
  remove_underscores = TRUE,
  remove_action_label = TRUE,
  label_mapping = NULL
)

Arguments

result_object: A list returned by `margot_causal_forest()`
model_name: Name of the model in the results to visualise
max_depth: Maximum depth of the tree (1L or 2L)
original_df: Optional dataframe with original data for showing untransformed values
x_padding: Horizontal padding for the plot (proportion)
y_padding: Vertical padding for the plot (proportion)
border_size: Size of node borders in lines
text_size: Size of text in plot elements
edge_label_offset: Offset for edge labels from connecting lines
span_ratio: Controls the aspect ratio of the plot
non_leaf_fill: Colour for non-leaf nodes (decision nodes)
title: Optional custom title for the plot
plot_margin: Margins around the plot
remove_tx_prefix: Whether to remove treatment prefixes from variable names
remove_z_suffix: Whether to remove z-suffixes from variable names
use_title_case: Whether to use title case for variable names
remove_underscores: Whether to replace underscores with spaces in variable names
remove_action_label: Whether to remove "Action:" prefix from leaf node labels
label_mapping: Optional list for renaming variables in the display

Value

A ggplot object with the decision tree visualisation