Beer by brewery

# Remove beers without abv listings; then see # of distinct beers breweries produce.
# Also compared to the mean/median # of beers made, but
# that doesn't seem particularly useful
beers_per_brewery <- reviews_raw %>%
  filter(!is.na(beer_abv))%>%
  # double count statement: first counts brewery, beer, # reviews
  # second aggregates that and returns number of distinct beers produced by brewery
  count(brewery_name, beer_name, sort = TRUE)%>%
  count(brewery_name, sort = TRUE)%>%
  # a count of how many brewerys only make one beer
  mutate(produce_one = sum(n==1)/n())

# create df of mean and median for easy plot labeling
brewery_measures <- beers_per_brewery %>%
  summarise(Mean = mean(n),
         Median = median(n))%>%
  gather()

# create plot
beers_per_brewery %>%
  ggplot(aes(n))+
  geom_histogram(fill = "#0072B2", color = 'white', alpha = .8)+
  geom_vline(data = brewery_measures,
             aes(xintercept = value, color = key),
             lty = 2,
             size = .9)+
  scale_x_log10()+
  scale_y_continuous(limits = c(0, 1100),
                     expand = c(0,0),
                     labels = comma_format())+
  theme(legend.position = c(.75, .87))+
  labs(x = "# of Beers",
       y = "# of Breweries",
       title = "Number of Unique Beers Produced by Breweries",
       subtitle = paste0(percent(beers_per_brewery$produce_one[1]), " of breweries in the dataset only produce 1 beer"),
       color = "")+
  my_theme_tweaks()