Fatal dog attacks by category of dog

fatal_dog_attacks.R

lapply(c("dplyr", "xml2", "rvest", "stringr", "ggplot2"), library, 
       character.only = TRUE)

url <- "https://en.wikipedia.org/wiki/Fatal_dog_attacks_in_the_United_States"

page <- read_html(url)

# Each year is represented by its own table
tbls <- page %>% 
  xml_find_all("//table") %>% 
  html_table()

# Extract the year for each table from their respective subheadings
years <- page %>% 
  xml_find_all("//h3/span[contains(text(), 'Fatalities reported in ')]") %>% 
  xml_text() %>% 
  str_replace_all("[^0-9]+", "") %>% 
  as.numeric() %>% 
  .[. != 1985] # No table for year 1985

# Add years to tables
for(i in seq_along(years)) {
  tbls[[i]]$year <- years[i]
}

# Bind together all tables
df <- rbind_all(tbls)

# Remove footnote references and plural endings
df$`Category of Dog` <- df$`Category of Dog` %>% 
  str_replace_all(" \\(.*\\)|\\(e?s\\)", "")

# Consolidate categories to include mixes
for(dog in c("Pit bull", "Rottweiler", "Bulldog", "Husky", "German Shepherd")) {
  df$`Category of Dog`[
    str_detect(df$`Category of Dog`, regex(dog, ignore_case = TRUE))] <- 
    paste(dog, " (incl. mixes)")
}

df %>% 
  group_by(`Category of Dog`) %>% 
  summarise(n = n()) %>% 
  filter(n > 2) %>% 
  ggplot(aes(x = reorder(`Category of Dog`, n), y = n)) + 
  geom_bar(stat = "identity") +
  coord_flip() +
  scale_y_continuous(expand = c(0, 0), limits = c(0, NA)) +
  theme_bw() +
  labs(x = NULL, y = "\nNumber of attacks", 
       title = "Fatal dog attacks by category of dog\n United States; 1887-2015")