一种简单的方法是使用expand
scale_y_continuous 的参数:
dt = Titanic %>% as.data.frame() %>%
filter(Survived == "Yes") %>%
mutate(FreqSurvived = ifelse(Freq > 100, Freq*1e+04,Freq))
ggplot(dt, aes(x = Age, y = FreqSurvived, fill = Sex)) +
geom_bar(stat = "identity", position = "dodge") +
facet_grid(Class ~ ., scales = "free") +
theme_bw() +
geom_text(aes(label = prettyNum(FreqSurvived,big.mark = ",")),
vjust = 0, position = position_dodge(0.9), size = 2) +
scale_y_continuous(expand = c(0.1,0))

使用的缺点expand
是它会在条形图的上方和下方增加空间。另一种方法是在图表上的条形上方高度绘制一些不可见的数据,这将迫使 ggplt 扩展轴范围以容纳这些虚拟数据。在这里,我添加了一些不可见的条,它们的高度是实际条的 1.2*:
Titanic %>% as.data.frame() %>%
filter(Survived == "Yes") %>%
mutate(FreqSurvived = ifelse(Freq > 100, Freq*1e+04,Freq)) %>%
ggplot( aes(x = Age, y = FreqSurvived, fill = Sex)) +
geom_bar(aes(y = FreqSurvived*1.2), stat = "identity",
position = "dodge", fill=NA) +
geom_bar(stat = "identity", position = "dodge") +
facet_grid(Class ~ ., scales = "free") +
theme_bw() +
geom_text(aes(label = prettyNum(FreqSurvived,big.mark = ",")),
vjust = 0,
position = position_dodge(0.9), size = 2)
