Week 15
setwd("~/Desktop/tidytuesday/data/2018/2018-07-10")
beers <- read_excel("week15_beers.xlsx", sheet = 1)
brewer <- read_excel("week15_beers.xlsx", sheet = 2)
beer <- inner_join(beers, brewer, by = c("brewery_id" = "id"))
byState <- beer %>%
group_by(state) %>%
summarise(numBrewer = length(unique(brewery_id)), count = n(), mabv = mean(abv, na.rm = T))
counties <- map_data("county")
state <- map_data("state")
stateInfo <- cbind.data.frame(abb = state.abb, name = tolower(state.name))
state <- inner_join(state, stateInfo, by = c("region" = "name"))
all_state <- inner_join(state, byState, by = c("abb" = "state"))
This palette isn’t very visually appealing, but in the spirit of beer, I’ll use it anyway.
# https://www.reddit.com/r/beer/comments/4gd24e/the_hex_colour_palette_of_beer/
beerPal <- c("#F3F993", "#F5F75C", "#F6F513", "#EAE615", "#E0D01B", "#D5BC26", "#CDAA37", "#C1963C", "#BE8C3A", "#BE823A", "#C17A37", "#BF7138", "#BC6733", "#B26033", "#A85839", "#985336", "#8D4C32", "#7C452D", "#6B3A1E", "#5D341A", "#4E2A0C", "#4A2727", "#361F1B", "#261716", "#231716", "#19100F", "#16100F", "#120D0C", "#100B0A", "#050B0A")
Where to Bar Crawl?
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = mabv), color = "grey") +
labs(fill = "mabv") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black") +
ggtitle("Where is the beer strongest?")
A stark (and believable) difference between Nevada and Utah.
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = numBrewer), color = "grey") +
labs(fill = "numBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black") +
ggtitle("Where are the most breweries?")
Colorado maintains it’s reputation.
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = count / numBrewer), color = "grey") +
labs(fill = "beerPerBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black") +
ggtitle("Where is the largest variety?")
Surprisingly Kansas is where it is at!
beer %>%
group_by(city, state) %>%
summarise(count = n(), numBrewer = length(unique(brewery_id))) %>%
arrange(desc(count))
# A tibble: 401 x 4
# Groups: city [384]
city state count numBrewer
<chr> <chr> <int> <int>
1 Grand Rapids MI 66 3
2 Chicago IL 55 9
3 Portland OR 52 11
4 Indianapolis IN 43 4
5 San Diego CA 42 8
6 Boulder CO 41 9
7 Denver CO 40 8
8 Brooklyn NY 38 4
9 Seattle WA 35 9
10 Longmont CO 33 1
# … with 391 more rowsbeer %>%
group_by(city, state) %>%
summarise(count = n(), numBrewer = length(unique(brewery_id))) %>%
arrange(desc(numBrewer))
# A tibble: 401 x 4
# Groups: city [384]
city state count numBrewer
<chr> <chr> <int> <int>
1 Portland OR 52 11
2 Boulder CO 41 9
3 Chicago IL 55 9
4 Seattle WA 35 9
5 Austin TX 25 8
6 Denver CO 40 8
7 San Diego CA 42 8
8 Bend OR 11 6
9 Portland ME 12 6
10 San Francisco CA 32 5
# … with 391 more rowsSomebody please tell me about the hidden gem of Grand Rapids. Apparently, it is Beer City, USA.
Styles
There are too many styles, so I pick some major ones and investigate them.
stout <- beer[str_detect(beer$style, "Stout"), ]
american <- beer[str_detect(beer$style, "American"), ]
ipa <- beer[str_detect(beer$style, "IPA"), ]
True American?
byStateA <- american %>%
group_by(state) %>%
summarise(numBrewer = length(unique(brewery_id)), count = n(), mabv = mean(abv, na.rm = T))
all_state <- inner_join(state, byStateA, by = c("abb" = "state"))
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = numBrewer), color = "grey") +
labs(fill = "numBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black")
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = count / numBrewer), color = "grey") +
labs(fill = "beerPerBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black")
Mississippi: American Beer’s Hearland
Representing my namesake
byStateS <- stout %>%
group_by(state) %>%
summarise(numBrewer = length(unique(brewery_id)), count = n(), mabv = mean(abv, na.rm = T))
all_state <- inner_join(state, byStateS, by = c("abb" = "state"))
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = numBrewer), color = "grey") +
labs(fill = "numBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black")
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = count / numBrewer), color = "grey") +
labs(fill = "beerPerBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black")
What’s up with some states having no stouts?!
The controversial IPA
byStateI <- ipa %>%
group_by(state) %>%
summarise(numBrewer = length(unique(brewery_id)), count = n(), mabv = mean(abv, na.rm = T))
all_state <- inner_join(state, byStateI, by = c("abb" = "state"))
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = numBrewer), color = "grey") +
labs(fill = "numBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black")
ggplot(data = state, mapping = aes(x = long, y = lat, group = group)) +
geom_polygon(data = all_state, aes(fill = count / numBrewer), color = "grey") +
labs(fill = "beerPerBrewer") +
scale_fill_gradientn(colors = beerPal) +
theme_void() +
geom_path(data = state, aes(x = long, y = lat, group = group), color = "black")
## what's up with Utah?
ut <- beer[which(beer$state == "UT"), ]
ut[str_detect(ut$style, "IPA"), ] ## double counting
# A tibble: 8 x 12
count.x abv ibu id name.x style brewery_id ounces count.y
<dbl> <dbl> <dbl> <dbl> <chr> <chr> <dbl> <dbl> <dbl>
1 1382 0.04 NA 644 Johnn… Amer… 399 16 400
2 2254 0.04 42 1925 Trade… Amer… 159 12 160
3 2255 0.073 83 1723 Hop N… Amer… 159 12 160
4 2258 0.073 82 1089 Hop N… Amer… 159 12 160
5 2300 0.09 75 1825 Squat… Amer… 302 12 303
6 2302 0.06 NA 1823 Wasat… Amer… 302 12 303
7 2303 0.06 NA 1682 Wasat… Amer… 302 12 303
8 2305 0.09 75 1680 Squat… Amer… 302 12 303
# … with 3 more variables: name.y <chr>, city <chr>, state <chr>West Virginia and Arkansas are not into IPAs.
Variation in ABV
Which styles have the most variation in alcohol content (of the top 20 most prevalent styles) given their average value?
beer %>%
group_by(style) %>%
summarise(count = n(), coeffVarabv = mean(abv, na.rm = T) / sd(abv, na.rm = T)) %>%
arrange(desc(count)) %>%
head(20) %>%
arrange(desc(coeffVarabv))
# A tibble: 20 x 3
style count coeffVarabv
<chr> <int> <dbl>
1 American Double / Imperial IPA 105 12.5
2 Kölsch 42 11.4
3 American Amber / Red Lager 29 10.7
4 American Blonde Ale 108 10.0
5 Märzen / Oktoberfest 30 9.72
6 Hefeweizen 40 9.44
7 American Pale Ale (APA) 245 8.63
8 Cider 37 7.58
9 American Pale Wheat Ale 97 7.55
10 American IPA 424 7.33
11 American Stout 39 7.17
12 German Pilsener 36 7.01
13 American Porter 68 6.96
14 American Amber / Red Ale 133 6.62
15 American Pale Lager 39 6.27
16 American Brown Ale 70 5.97
17 American Black Ale 36 5.49
18 Fruit / Vegetable Beer 49 5.44
19 Saison / Farmhouse Ale 52 5.41
20 Witbier 51 4.79Fancy string matching for another time: match the beer style to the colors listed here.