In joemarlo/plotBart: Diagnostic and Plotting Functions to Supplement 'bartCause'

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  message = FALSE,
  warning = FALSE
  # dev = "png", # uncomment if issues running locally on mac
  # dev.args = list(type = "cairo-png")
)

plotBart uses ggplot2 to construct plots. These functions return ggplot objects that can be manipulated similar to standard ggplot2::ggplot() calls. Plot titles, subtitles, labels, captions, and themes can be easily customized using ggplot syntax.

library(plotBart)
data(lalonde)
confounders <- c('age', 'educ', 'black', 'hisp', 'married', 'nodegr')

# plot balance across treatment and control groups
p <- plot_balance(.data = lalonde,
                  treatment = 'treat',
                  confounders = confounders, 
                  compare = 'means', 
                  estimand = 'ATE')
p

Modify the plot labels:

p +
  labs(title = 'My comments on the results',
       subtitle = NULL,
       caption = 'December 2021',
       x = 'Mean diff b/t treatment and control')

Or change the theme entirely:

p + 
  theme_classic()
# set the theme for all plots within this R session
theme_set(theme_bw())

Extract data to reconstruct the plot:

p$data

Refer to the ggplot2 documentation for more details:

• ggplot2 cheatsheet
• Labels
• Complete themes
• Modify a theme

joemarlo/plotBart documentation built on May 31, 2024, 12:22 p.m.