创建按多个 x 参数分组的堆积条形图

Question

我在创建堆叠条形图时遇到困难，其中 x 轴有多个分组。

 df <- data.frame(random = c(1, 1, 1, 1, 1, 1, 1),
                  value = c("s1", "s1", "s1", "s1", "s1", "s2", "s3"),
                  countries = factor(c("US", "Japan", "UK", "Germany", "France", "France", "France")),
                  week = factor(c("week1", "week2", "week1", "week2", "week1", "week1", "week2")))

在此示例中，我尝试按 week 和 countries 对 x 轴进行分组。 我想通过按 week、countries 和 value 分组来显示 value 出现次数的百分比。我希望条形图的输出显示为：

• 美国，第 1 周：100%
• 日本，第 2 周：100%
• 英国，第 1 周：100%
• 德国，第 1 周：100%
• 法国，第 1 周：50%（value-s1），50%（value-s2）
• 法国，第 2 周：100%（价值-s3）

我当前的代码是：

  result <- df %>% group_by_(df$countries, df$week, df$value) %>% summarise(total = n()) #get number of occurences
  ggplot(result, aes(x = interaction(result[["countries"]], result[["week"]]), y = result[["total"]], fill = result[["total"]])) +
    geom_bar(position = "fill", stat = "identity") +
    scale_y_continuous(labels = scales::percent_format())

你能帮我解决这个逻辑吗？我相信 scale_y_continuous(labels = scales::percent_format()) 应该为我计算百分比。

Answer 1

不确定填充变量应该是什么。

result <- df %>% 
  unite("group", countries:week, sep = ", ") %>%
  count(group, value) %>% 
  group_by(group) %>%
  mutate(p = round(n / sum(n), 2))

ggplot(result, aes(x = group, y = p, fill = as.factor(value))) +
  geom_bar(position = "fill", stat = "identity") +
  scale_y_continuous(labels = scales::percent_format()) +
  coord_flip()

编辑

如果我理解正确，请尝试：

result <- df %>% 
  unite("group", countries:week, sep = ", ") %>%
  count(group, value) %>% 
  group_by(group) %>%
  mutate(p = round(n / sum(n), 2)) %>%
  separate("group", c("country", "week"), sep = ", ")

ggplot(result, aes(x = week, y = p, fill = as.factor(value))) +
  geom_bar(position = "fill", stat = "identity", width = 0.5) +
  facet_grid(~country,scales = "free_x") +
  scale_y_continuous(labels = scales::percent_format())