Lollypop Plots

Mean Wage, by Sex and Sector

From the CPS85 data set in the mosaicData package:

library(dplyr)

MeanWages <-
  mosaicData::CPS85 %>% 
  group_by(sector, sex) %>% 
  summarise(meanWage = mean(wage, na.rm = TRUE))

Here’s the table:

The Lollypop Plot: a Graphical Approach

There isn’t much data in the table, so if we insist on making a graph it should have a very light footprint. Here’s one that I call a lollypop plot:

library(ggplot2)

MeanWages %>% 
  ggplot() +
  geom_point(aes(x = sector,
                y = meanWage,
                color = sex),
             size = 2.5,
             position = position_dodge(width = 0.7)) +
  geom_bar(aes(x = sector, y = meanWage, fill = sex),
           width = 0.1, stat = "identity",
           position = position_dodge(width = 0.7)) +
  labs(x = "Sector of Employment",
       y = "mean wage ($/hr)",
       title = "Males made more than females\nin almost every sector!") +
  theme(panel.grid.major.x = element_blank())

Priority

I discovered that lollypop plots have already been investigated in ggplot2 by Yan Holtz. See his excellent R Graph Gallery.

Modufying Yan’s ideas a bit, we might create a double-pop lollypop plot as follows:

MeanWages %>% 
  ggplot(aes(x = sector, y = meanWage, group = sector)) +
  geom_point(aes(color = sex), size = 2) +
  scale_color_manual(values = c("blue", "red")) +
  labs(x = NULL, y = "mean wage ($/hr)") +
  geom_line() +
  coord_flip() +
  theme_light() +
  theme(panel.border = element_blank())

Let’s consider re-ordering the sectors by the value male mean wage:

MeanWages %>% 
  inner_join(MeanWages %>% 
               ungroup() %>% 
               filter(sex == "M") %>% 
               rename(maleMean = meanWage) %>% 
               select(-sex),
             by = "sector") %>% 
  ggplot(aes(x = forcats::fct_reorder(sector, maleMean),
             y = meanWage, group = sector)) +
  geom_point(aes(color = sex), size = 2) +
  scale_color_manual(values = c("blue", "red")) +
  geom_line() +
  labs(x = NULL, y = "mean wage ($/hr)") +
  coord_flip() +
  theme_light() +
  theme(panel.border = element_blank())