Basic code for creating a satellite diagram (aka visual ANOVA)

satellite_viz_anova.R

# School and regional data (% of students earning 2+ on Mock AP exams)
schools_plot_df %>% 
  ggplot(aes(y=region, x = pct_2plus)) +
    # plot the schools adn regions first
    geom_point(aes(size = n_stus,
                   color = str_detect(school, "Average")), #Name of region = "Region Name Average" is in school field
               show.legend = FALSE,
               alpha = 0.75  #lighten up the points so we can see overlaps better
            ) +
  
  # curve and text pointing out school average example
  annotate("curve", x=.15, y=2.25, xend = .19, yend = 2.05,
           color = idea_colors$magenta,
           curvature = -.2, arrow = arrow(length = unit(2, "mm")))+
  
  annotate("text", x=.15, y=2.25, hjust = 1,
           label = "One school",
           color = idea_colors$magenta) +
  
   # curve and text pointing out region average example
  annotate("curve", x=.3, y=2.3, xend = .265, yend = 2.075,
           color = idea_colors$darkblue,
           curvature = .2, arrow = arrow(length = unit(2, "mm")))+
  
  annotate("text", x=.30, y=2.3, hjust = -0.01,
           label = "Region's average",
           color = idea_colors$darkblue) +

  scale_color_idea(reverse = TRUE) +
  scale_x_continuous(labels = scales::percent_format(), expand =expansion(mult = c(0, .1))) +
  theme_idea_light() +
  theme(axis.text.y = element_text(color = school_label_colors$color)) +
  labs(y = "",
       x = "% of AP Mock Exams earning 2+ points",
       title = "Percent of Spring 20201 AP Mock Exams earning 2+ points",
       caption = "Source: IDEA internal assessments") +
  # + NULL is a trick to make it easy to comment out lines with out a + causing things to run over.
  NULL