How to display values of a column inside bars of a bar chart

Using the code below, I could create the below graph. I would like to make some customization as follows:

1- drop the numbers from the beginning of labels in the legend (e.g.,4.Male NCG to Male NCG) but do not change the order of values

2- display the values of meanHLE_ inside the bars and define using a new legend

3-separate male and female categories in the graph

#my data

sample_label<-c("4.Male NCG","4.Male NCG","4.Male NCG",
                    "5.Male caregivers<14h/w","5.Male caregivers<14h/w",
                    "5.Male caregivers<14h/w","6.Male caregivers>=14h/w",
                    "6.Male caregivers>=14h/w","6.Male caregivers>=14h/w",
                    "7.Female NCG","7.Female NCG","7.Female NCG", 
                    "8.Female caregivers<14h/w", "8.Female caregivers<14h/w",
                    "8.Female caregivers<14h/w", "9.Female caregivers>=14h/w",
                    "9.Female caregivers>=14h/w","9.Female caregivers>=14h/w")

        Age_Group_<-c("50-51","60-61","70-71","50-51","60-61","70-71",
                      "50-51","60-61","70-71","50-51","60-61","70-71",
                      "50-51","60-61","70-71","50-51","60-61","70-71")
        meanTLE_<-c(32.4,24,16.3,34.4,25.6,17.3,33.4,24.7,16.8,
                    36,26.8,18.1,37.3,27.9,18.8,36.6,27.3,18.5)
        meanHLE_<-c(24.8,18.3,12.3,27.2,20.2,13.6,25.3,18.7,12.6,
                    28.8,21.4,14.4,30.7,22.9,15.4,29.1,21.6,14.5)

2.plot section

gender<-data.frame(sample_label,Age_Group_,meanTLE_,meanHLE_)
ggplot(gender, aes(x =Age_Group_, y = meanTLE_, fill=sample_label)) + geom_bar(stat ="identity", position = "dodge2") + #fill = "#B61E2E"
  geom_text(
    aes(label = meanTLE_),
    vjust = 0,
    colour = "black", 
    position = position_dodge(width=0.9),
    fontface = "bold",
    size=3,
    angle = 90,
    hjust = 0
  ) +ylim(0,50)+

  labs(
    x = "Age Groups",
    y = "Total Life Expactacny",
    face = "bold"
  ) +
  # coord_flip() +
  theme_bw() +
  # scale_fill_manual(values=c("meanHLE_")) + 
  theme(legend.title=element_blank(),legend.text = element_text(face = "bold"),plot.title = element_text(
    hjust = 0.5,
    size = 15,
    colour = "Black",
    face = "bold"
  ),
  plot.caption = element_text(hjust = 0, color = "black", face = "bold", size=12.5))

Solution

I think the plot below does most of what you indicated in your OP and comments. However, I think this plot is very busy and the patterns aren't as transparent as they could be, so I've added a couple of other options.

library(tidyverse)

pd = position_dodge(0.9)

gender %>% 
  mutate(sex=str_extract(sample_label, "Male|Female"),
         sample_label=gsub(".*ale ", "", sample_label),
         sample_label=fct_relevel(sample_label, "NCG")) %>% 
  ggplot(aes(x =Age_Group_, y = meanTLE_, fill=sample_label, group=sample_label)) + 
  # Dodge value labels and bars by same amount
  geom_col(position = pd, width=0.85) + 
  geom_text(aes(label=sprintf("%1.1f", meanTLE_)), hjust=0,
            colour = "black", fontface = "bold", size=3, angle = 90,
            # Dodge value labels and bars by same amount
            position = pd) + 
  geom_col(aes(y=meanHLE_), width=0.4, size=0.2, colour="grey50", fill="white", position=pd) +
  geom_text(aes(label=sprintf("%1.1f", meanHLE_), 
                y=meanHLE_, colour=sample_label), position=pd,
            colour = "black", fontface = "bold", size=3, angle = 90,) +
  geom_text(data=. %>% 
              group_by(sex) %>%
              filter(Age_Group_=="70-71", grepl(">=14", sample_label)) %>% 
              ungroup %>% 
              pivot_longer(starts_with("mean")), hjust=0, colour="grey30", size=3,
            aes(x=3.5, y=value, label=gsub("mean(.*)_", "\\1", name))) +
  facet_grid(cols=vars(sex)) +
  scale_y_continuous(limits=c(0,41), expand=c(0,0)) +
  expand_limits(x=3.9) +
  scale_colour_manual(values=hcl(seq(15,375,length=4)[1:3], 100, 80)) +
  labs(x = "Age Groups", y = "Total Life Expectancy") +
  theme_bw() +
  theme(legend.title=element_blank(),
        legend.text = element_text(face = "bold"),
        plot.title = element_text(hjust = 0.5, size = 15, colour = "Black", face = "bold"),
        plot.caption = element_text(hjust = 0, color = "black", face = "bold", size=12.5))

The plot below gets rid of the bars and uses values as point markers. I've reshaped the data to long format so that we need only a single call to each geom to plot the values and we can generate a legend with a color mapping. I've added subtle lines just to guide the eye. If you want to remove them, set size=0 in geom_line. Don't get rid of the geom_line call, as it's necessary in order to generate the legend.

gender %>% 
  mutate(sex=str_extract(sample_label, "Male|Female"),
         sample_label=gsub(".*ale ", "", sample_label)) %>% 
  pivot_longer(cols=starts_with("mean")) %>% 
  ggplot(aes(y=sample_label, x=value, colour=name, group=name)) + 
    geom_line(size=0.4, alpha=0.5) +
    geom_text(aes(label=sprintf("%1.1f", value)), 
              fontface = "bold", size=3, show.legend=FALSE) + 
    facet_grid(rows=vars(sex, Age_Group_)) +
    scale_x_continuous(limits=c(0,40), expand=c(0,0)) +
    scale_colour_discrete(labels=c("Healthy", "Total")) +
    labs(x = "Mean (years)",
         title="Life expectancy by age, sex, and caregiver status") +
    theme_bw() +
    theme(legend.title=element_blank(),
          axis.title.y=element_blank(),
          legend.text = element_text(face = "bold"),
          plot.title = element_text(hjust = 0.5, size = 15, face = "bold"),
          plot.caption = element_text(hjust=0, color="black", face="bold", size=12.5),
          strip.text.y=element_text(angle=0),
          legend.position="bottom",
          legend.box.margin=margin(t=-10)) +
    guides(colour=guide_legend(override.aes=list(size=3, alpha=1)))

This version seems to me to be the most transparent regarding the key patterns in the data. The data value text labels overlap, so maybe dodging or using point markers would be more effective.

gender %>% 
  mutate(sex=str_extract(sample_label, "Male|Female"),
         sample_label=gsub(".*ale ", "", sample_label)) %>% 
  pivot_longer(cols=starts_with("mean")) %>% 
  ggplot(aes(y=value, x=Age_Group_, colour=sample_label)) + 
    geom_line(aes(linetype=name, group=interaction(name, sample_label)), 
              size=0.6, alpha=0.3) +
    geom_text(aes(label=sprintf("%1.1f", value)), size=3, show.legend=FALSE) + 
    facet_grid(cols=vars(sex)) +
    scale_y_continuous(limits=c(0,40), expand=c(0,0)) +
    scale_linetype_discrete(labels=c("Healthy", "Total")) +
    labs(x = "Age Group", y="years", 
         colour="Caregiver Status", linetype="Life Expectancy Type",
         title="Life expectancy by age, sex, and caregiver status") +
    theme_bw() +
    theme(legend.text = element_text(face = "bold"),
          plot.title = element_text(hjust = 0.5, size = 15, face = "bold"),
          plot.caption = element_text(hjust = 0, color = "black", face = "bold", size=12.5)) +
    guides(linetype=guide_legend(reverse=TRUE, override.aes=list(size=1)),
           colour=guide_legend(override.aes=list(size=1, alpha=1)))