Population pyramids

A population pyramid visualises the age-gender distribution. This recipe uses the latest mid-year population estimates for a local authority and visualises them by gender and 5-year age bands.

Ingredients

Data sources

Nomis, Office for National Statistics

R packages

Instructions

Load the necessary R packages.

library(tidyverse) ; library(ggpol)

Visit Nomis and select your chosen Geography, Date, Age and Sex. Here we’ve chosen Trafford (Geography), 2018 (Date), individual ages (Age) and both genders (Sex).
Navigate to ‘Format / Layout’ and choose ‘Nomis API’ as your format. Punch the ‘Download Data’ button. Select the ‘Tabulation links’ tab and right-click the ‘Comma Separated Values (csv)’ file to obtain the URL path.
Paste the URL in the path argument of read_csv.

df <- read_csv("http://www.nomisweb.co.uk/api/v01/dataset/NM_2002_1.data.csv?geography=E08000009&date=latest&gender=1,2&c_age=101...191&measures=20100&select=date_name,geography_name,geography_code,gender_name,c_age_name,measures_name,obs_value,obs_status_name")

Prepare the data for visualisation. Here we group the individual ages into bands and calculate the population size for each gender.

population <- df %>% 
  select(gender = GENDER_NAME, age = C_AGE_NAME, n = OBS_VALUE) %>% 
  mutate(gender = factor(gender, levels = c("Male", "Female")),
         age = parse_number(age),
         ageband = cut(age,
                       breaks = c(0,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80,85,90,120),
                       labels = c("0-4","5-9","10-14","15-19","20-24","25-29","30-34","35-39",
                                  "40-44","45-49","50-54","55-59","60-64","65-69","70-74",
                                  "75-79","80-84","85-89","90+"),
                       right = FALSE)) %>% 
  group_by(gender, ageband) %>% 
  summarise(n = sum(n)) %>% 
  mutate(n = case_when(gender == "Male" ~ n * -1, TRUE ~ n))

Plot the population pyramid.

ggplot(population, aes(x = ageband, y = n, fill = gender)) +
  geom_col() + 
  scale_fill_manual(values = c("#7FC5DC", "#7FDCC5"), labels = c("Female", "Male")) +
  facet_share(~gender, dir = "h", scales = "free", reverse_num = TRUE) +
  coord_flip() +
  labs(x = NULL, y = NULL, 
       title = "Age composition of Trafford in mid-2018",
       caption = "Source: Office for National Statistics", 
       fill = NULL) +
  theme_minimal() +
  theme(plot.margin = unit(c(1,1,1,1), "cm"),
        panel.spacing = unit(0.05, "lines"),
        panel.grid.major.y = element_blank(),
        plot.title = element_text(size = 13, face = "bold", hjust = 0.5, vjust = 4),
        strip.text = element_text(size = 11, vjust = 1),
        axis.text.x = element_text(size = 9),
        axis.text.y = element_text(size = 9),
        plot.caption = element_text(size = 9, color = "grey50", hjust = 1, margin = margin(t = 15)),
        legend.position = "none")

Output the chart as a PNG file.

ggsave("population_pyramids.png", dpi = 300)

Supplementary code

If you want to compare the age distribution of your chosen area with the total population of England you can use this code.

library(tidyverse) ; library(ggpol)

df <- read_csv("http://www.nomisweb.co.uk/api/v01/dataset/NM_2002_1.data.csv?geography=2092957699,E08000009&date=latest&gender=1,2&c_age=101...191&measures=20100&select=date_name,geography_name,geography_code,gender_name,c_age_name,measures_name,obs_value,obs_status_name")

population <- df %>% 
  select(area_name = GEOGRAPHY_NAME, gender = GENDER_NAME, age = C_AGE_NAME, n = OBS_VALUE) %>% 
  mutate(gender = factor(gender, levels = c("Male", "Female")),
         age = parse_number(age),
         ageband = cut(age,
                       breaks = c(0,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80,85,90,120),
                       labels = c("0-4","5-9","10-14","15-19","20-24","25-29","30-34","35-39",
                                  "40-44","45-49","50-54","55-59","60-64","65-69","70-74",
                                  "75-79","80-84","85-89","90+"),
                       right = FALSE)) %>% 
  group_by(area_name, gender, ageband) %>% 
  summarise(n = sum(n)) %>%
  group_by(area_name) %>% 
  mutate(percent = round(n/sum(n)*100, 1),
         percent = case_when(gender == "Male" ~ percent * -1, TRUE ~ percent))

ggplot() +
  geom_col(data = filter(population, area_name == "Trafford"), 
           aes(x = ageband, y = percent, fill = gender, alpha = 0.5)) + 
  geom_line(data = filter(population, area_name == "England"), 
            aes(x = ageband, y = percent, group = gender, colour = gender), stat = "identity", size = 1) +
  scale_fill_manual(values = c("#7FC5DC", "#7FDCC5"), labels = c("Female", "Male")) +
  scale_colour_manual(values = c("#7FC5DC", "#7FDCC5"), labels = c("Female", "Male")) +
  facet_share(~gender, dir = "h", scales = "free", reverse_num = TRUE) +
  coord_flip() +
  labs(x = NULL, y = "%", 
       title = "Age composition of Trafford compared with England in mid-2018",
       caption = "Source: Office for National Statistics", 
       fill = NULL) +
  theme_minimal() +
  theme(plot.margin = unit(c(1,1,1,1), "cm"),
        panel.spacing = unit(0.05, "lines"),
        panel.grid.major.y = element_blank(),
        plot.title = element_text(size = 13, face = "bold", hjust = 0.5, vjust = 4),
        strip.text = element_text(size = 11, vjust = 1),
        axis.text.x = element_text(size = 9),
        axis.text.y = element_text(size = 9),
        plot.caption = element_text(size = 9, color = "grey50", hjust = 1, margin = margin(t = 15)),
        legend.position = "none")