What's the most popular car in the UK?

The Office of National Statistics is a brilliant source of interesting data sets. This page lists recently released datasets. Check regularly and you’ll soon notice these data sets feed into news articles by the Economist, BBC news and more.

One I recently thought to explore detailed the count of car makes and models in the UK. I wonder what’s the most popular car make and model in the UK?

library(tidyverse)
library(readxl)
library(stringr)

path <- "~/Learning/R/Datasets/vehicle-statistics-make-model.xlsx"
car_data <- read_excel(path = path, skip = 8, col_names = FALSE, sheet =  1)

Great, we’ve got our data. But what does it look like?

oh no.

It’s not as we’d like it, but we can certainly handle it.

We’ll need to tidy it up a little before the real fun begins.

names <- read_excel(path = path, skip = 6, col_names = F, n_max = 1)
names[2] <- "Model"

names(car_data) <- names

tidy_cars <- car_data %>%
  gather(key = Period, value = Count, -Make, -Model) %>%
  extract(Period, c("Year", "Quarter", "Region"), "(\\d{4}) Q(1|2|3|4) (UK|GB)") %>%
  mutate(Year = as.integer(Year),
         Quarter = as.integer(Quarter),
         Make = str_trim(Make),
         Model = str_trim(Model),
         YQ = lubridate::yq(paste(Year,"/",Quarter))) %>%
  filter(stringr::str_length(Make) < 20,
         !is.na(Count))

Much better!

Now that we have our tidy dataset, lets explore the latest counts in the UK, from Q2 2017.

tidy_cars %>% 
  filter(Year == 2017, Quarter == 2, Region == "UK") %>%
  count(Make, wt = Count) %>%
  arrange(desc(n)) %>%
  top_n(n, n = 30) %>%
  ggplot(aes(reorder(Make, n), n)) + 
  geom_bar(stat = "identity") + 
  labs(x = "Make", y = "Count", 
       title = "The 30 most popular car makes in the UK as of Q2 2017", 
       subtitle = "Ford, Vauxhal and Volkswagen are significantly more popular than other makes") +
  coord_flip() + scale_y_continuous(labels = scales::comma)

tidy_cars %>% 
  select(Make, Model) %>%
  filter(Make == "FORD", grepl("B-MAX", Model)) %>%
  distinct() %>% 
  knitr::kable()

# helper function to split Model on the first " " and extract the first chunk
simplify_model <- function(model) {
  split_list <- stringr::str_split(model, " ") %>% `[[`(1)
  
  split_list[[1]]
}

tidy_cars %>% 
  filter(Year == 2017, Quarter == 2, Region == "UK") %>%
  mutate(simple_model = map_chr(Model, ~ simplify_model(.x))) %>%
  group_by(Make, simple_model) %>%
  summarise(Count = sum(Count)) %>% 
  ungroup() %>%
  top_n(25, Count) %>%
  ggplot(aes(x = (reorder(simple_model, Count)), y = Count, fill = Make)) + 
    geom_bar(stat = "identity") + 
    scale_y_continuous(labels = scales::comma) +
    labs(x = "Model",
         y = "Count", 
         title = "The most popular car makes as of Q2 2017", 
         subtitle = "Ford, Vauxhall and Volkswagen are the most popular makes") +
    coord_flip()