Working with aggregated data - Creating a barplot

[Eidenai]

Hello R-Community,

this is my first post and I try to be as precise as possible. Please don´t bother to ask for more information.

This is the case:
I got some aggregated data, containing city names, people per city, and share of females per city. I´d like to create a barplot, displaying the proportion of males and females per city. Also I´d like to have the share of females per city displayed in descending order.
The result should look something like this: (Picture taken form Hadley Wickhams - R for Data Science)

I figured, using the position = "fill" command would be a good idea. This is how far i got:

Creating and mutating the data:

library(tidyverse)

City <- c("Barcelona", "Bari", "Bologna", "Borgholm", "Duesseldorf", "Dublin", "Edinburgh", "Florence", "Frankfurt", "Genoa",
          "Hamburg", "Karvina", "Krakow", "Milan", "Most", "Ostrava", "Palermo", "Prague", "Rome", "Saarbruecken", "Stockholm",
          "Teplice", "Turin", "Warsaw", "Bochum", "London", "Berlin", "Paris", "Lissboa", "Vienna", "Copenhagen", "Helsinki",
          "Riga", "Madrid")

People <- c(472, 60, 657, 90, 389, 49, 120, 269, 763, 131, 60, 120, 78, 70, 202, 128, 184, 93, 120, 250, 79, 105, 101, 412,
            102, 87, 99, 69, 410, 312, 210, 100, 65, 79)

Share_of_Females <- c(0.3827029, 0.2649492, 0.5627421, 0.3717596, 0.5583374, 0.4967123, 0.3787804, 0.4668641, 0.4715904, 0.5424107,
                      0.4986961, 0.5938898, 0.6258508, 0.5458969, 0.6113255, 0.5897642, 0.6166618, 0.5848115, 0.6146192, 0.4975600,
                      0.4061146, 0.5035726, 0.5511435, 0.5651329, 0.6123814, 0.6213054, 0.6113255, 0.6315025, 0.6166618,
                      0.4011456, 0.4975600, 0.4013462, 0.4975600, 0.4011456)


df <- data.frame(City, People, Share_of_Females)

df <- mutate(df,
             Share_of_Males = 1 - Share_of_Females,
             males = round(People * Share_of_Males),
             females = round(People * Share_of_Females))

Since this data is aggregated, I wanted to get a new data.frame, where males and females are displayed in one column by gender.

mcity <- data.frame(City = df$City, gender = "m", n = df$males, People = df$People, share = df$Share_of_Males)
fcity <- data.frame(City = df$City, gender = "f", n = df$females, People = df$People, share = df$Share_of_Females)

df.plot <- rbind(mcity, fcity)

df.plot %>% 
  ggplot() +
  geom_bar(mapping = aes(x = City, y = n , fill = gender), stat = "identity", position = "fill") +
  labs(x = "City",
       y = "Relative Proportion Male vs. Female",
       title = "Relative Proportion Male vs. Female per City") +
  coord_flip()

Do you have any ideas on how I can display the share of females per city in descending order?

Thanks a lot in advance!

[FJCC]

I bet there is a better way to do this but this works.

library(ggplot2)
library(dplyr)


City <- c("Barcelona", "Bari", "Bologna", "Borgholm", "Duesseldorf", "Dublin", "Edinburgh", "Florence", "Frankfurt", "Genoa",
          "Hamburg", "Karvina", "Krakow", "Milan", "Most", "Ostrava", "Palermo", "Prague", "Rome", "Saarbruecken", "Stockholm",
          "Teplice", "Turin", "Warsaw", "Bochum", "London", "Berlin", "Paris", "Lissboa", "Vienna", "Copenhagen", "Helsinki",
          "Riga", "Madrid")

People <- c(472, 60, 657, 90, 389, 49, 120, 269, 763, 131, 60, 120, 78, 70, 202, 128, 184, 93, 120, 250, 79, 105, 101, 412,
            102, 87, 99, 69, 410, 312, 210, 100, 65, 79)

Share_of_Females <- c(0.3827029, 0.2649492, 0.5627421, 0.3717596, 0.5583374, 0.4967123, 0.3787804, 0.4668641, 0.4715904, 0.5424107,
                      0.4986961, 0.5938898, 0.6258508, 0.5458969, 0.6113255, 0.5897642, 0.6166618, 0.5848115, 0.6146192, 0.4975600,
                      0.4061146, 0.5035726, 0.5511435, 0.5651329, 0.6123814, 0.6213054, 0.6113255, 0.6315025, 0.6166618,
                      0.4011456, 0.4975600, 0.4013462, 0.4975600, 0.4011456)


df <- data.frame(City, People, Share_of_Females, stringsAsFactors = FALSE)

df <- mutate(df,
             Share_of_Males = 1 - Share_of_Females,
             males = round(People * Share_of_Males),
             females = round(People * Share_of_Females))

mcity <- data.frame(City = df$City, gender = "m", n = df$males, People = df$People, 
                    share = df$Share_of_Males, stringsAsFactors = FALSE)
fcity <- data.frame(City = df$City, gender = "f", n = df$females, People = df$People, 
                    share = df$Share_of_Females, stringsAsFactors = FALSE)
fcity$City <- factor(fcity$City, levels = fcity$City[order(fcity$share)])

mcity$City <- factor(mcity$City, levels = levels(fcity$City))

df.plot <- rbind(mcity, fcity)

df.plot %>% 
  ggplot() +
  geom_bar(mapping = aes(x = City, y = n , fill = gender), stat = "identity", position = "fill") +
  labs(x = "City",
       y = "Relative Proportion Male vs. Female",
       title = "Relative Proportion Male vs. Female per City") +
  coord_flip()

^{Created on 2020-05-20 by the reprex package (v0.2.1)}

[Eidenai]

Thank you very much for the fast reply! This is exactly what I was looking for.

As I see this now I recognize that some cities don´t quite fit the criteria of being displayed in descending order.

For example Bochum vs. Berlin in the upper part or Helsinki vs. Madrid in the lower part of the plot.

For my case this plot works pretty fine! I´m just curious about the reason.

Thanks again.