如何突出显示ggplot2中的列

Question

我有下面的图，我想突出显示两个西瓜的列，因为它的juice_content和重量最高。我知道如何更改列的颜色，但我想突出显示整个列。关于如何实现这一点的任何想法？网上似乎没有任何类似的内容。

fruits <- c("apple","orange","watermelons")
juice_content <- c(10,1,1000)
weight <- c(5,2,2000)
df <- data.frame(fruits,juice_content,weight)
df <- gather(df,compare,measure,juice_content:weight, factor_key=TRUE)
plot <- ggplot(df, aes(fruits,measure, fill=compare)) + geom_bar(stat="identity", position=position_dodge()) + scale_y_log10()

Answer 1

一种选择是使用gghighlight

library(gghighlight)
ggplot(df, aes(fruits,measure, fill = compare)) +
    geom_col(position = position_dodge()) +
    scale_y_log10() +
    gghighlight(fruits == "watermelons")

---

根据您的评论，如何使用不同的alpha值

ggplot(df, aes(fruits,measure)) +
    geom_col(data = . %>% filter(fruits == "watermelons"),
        mapping = aes(fill = compare),
        position = position_dodge()) +
    geom_col(data = . %>% filter(fruits != "watermelons"),
        mapping = aes(fill = compare),
        alpha = 0.2,
        position = position_dodge()) +
    scale_y_log10()

---

或者您也可以使用一个geom_col和一个有条件的alpha（感谢@Tjebo）来实现相同的目标

ggplot(df, aes(fruits, measure)) +
    geom_col(
        mapping = aes(fill = compare, alpha = fruits == 'watermelons'),
        position = position_dodge()) +
    scale_alpha_manual(values = c(0.2, 1)) +
    scale_y_log10()

Answer 2

您可以使用geom_area突出显示栏后面。您必须先将x比例尺强制设为离散，这就是为什么我使用geom_blank（请参见答案geom_ribbon overlay when x-axis is discrete）的原因，并指出geom_ribbon和geom_area实际上是相同的，除了geom_area的ymin始终为0

#minor edit so that the level isn't hard coded
 watermelon_level <- which(levels(df$fruits) == "watermelons")

AreaDF <- data.frame(fruits = c(watermelon_level-0.5,watermelon_level+0.5))


plot <- ggplot(df, aes(fruits)) + 
  geom_blank(aes(y=measure, fill=compare))+
  geom_area(data = AreaDF, aes( y = max(df$measure)), fill= "yellow")+
  geom_bar(aes(y=measure, fill=compare),stat="identity", position=position_dodge()) + scale_y_log10()

编辑以发表评论

如果要突出显示多个水果，则可以执行以下操作。您需要一个data.frame，其中要包含geom_area x和y的位置，包括将它们之间的距离降低到0。我敢肯定，获取data.frame的方法略微一些，但这是可行的

highlight_level <- which(levels(df$fruits) %in% c("apple", "watermelons"))

AreaDF <- data.frame(fruits = unlist(lapply(highlight_level, function(x) c(x -0.51,x -0.5,x+0.5,x+0.51))),
                     yval = rep(c(1,max(df$measure),max(df$measure),1), length(highlight_level)))




AreaDF <- AreaDF %>% mutate(
  yval = ifelse(floor(fruits) %in% highlight_level & ceiling(fruits) %in% highlight_level, max(df$measure), yval)) %>%
  arrange(fruits) %>% distinct()


plot <- ggplot(df, aes(fruits)) + 
  geom_blank(aes(y=measure, fill=compare))+
  geom_area(data = AreaDF, aes(y = yval ), fill= "yellow")+
  geom_bar(aes(y=measure, fill=compare),stat="identity", position=position_dodge()) + scale_y_log10()
plot