问题描述
我想绘制一条带有密度曲线的直方图，然后在顶部边缘上方放置一个箱形图。我知道如何使用 hist（）， boxplot（）和 layout（ ）函数，或使用 ggplot2 和 grid 包中的函数。但是，我正在寻找在 ggplot2 和 ggMarginal（）函数的特定解决方案> ggExtra 软件包。在提出问题之前，让我们模拟一些数据：
I would like to draw a histogram with a density curve and then put a boxplot above the top margin. I know how to do this using the hist(), boxplot() and layout() functions, or using functions from the ggplot2 and grid packages. However, I am looking for a specific solution using ggplot2 and the ggMarginal() function within the ggExtra package. Let's simulate some data before I present my problem:
library(ggplot2)
library(ggExtra)
set.seed(1234)
vdat = data.frame(V1 = c(sample(1:10, 100, T), 99))
vname = colnames(vdat)[1]
boxplot(vdat[[vname]], horizontal = T)
注意，我明确插入离群值99到1到10的数字样本中。因此，当我绘制箱线图时，应将99显示为离群值。
To note, I explicitly insert an outlier 99 into a sample of numbers from 1 to 10. Hence, when I draw the boxplot, 99 should be displayed as an outlier.
我可以使用 ggplot2 轻松绘制直方图。
I can easily draw a histogram using ggplot2.
p = ggplot(data=vdat, aes_string(x=vname)) +
geom_histogram(aes(y=stat(density)),
bins=nclass.Sturges(vdat[[vname]])+1,
color="black", fill="steelblue", na.rm=T) +
geom_density(na.rm=T) +
theme_bw()
p
当我尝试使用 ggMarginal 添加一个边缘盒图，添加的盒图不正确。
When I try to use ggMarginal to add a marginal boxplot, the added boxplots are not right.
p1 = ggMarginal(p, type="boxplot")
p1
右边的方框图可能是正确的。但是最上面的那个，我需要的那个，绝对是错误的。离群值99不存在，中位数显然不正确。
The boxplot on the right might be right. But the one on top, which is the very one I need, is definitely wrong. The outlier 99 is not there and the median is clearly not right.
当我尝试不提供 p1 时，却提供了原始数据 x 和帮助文档建议的 y ，我得到了正确的箱线图，但直方图现在不见了。
When I try not to provide p1, but the original data, x, and y as suggested by the help documentation, I get the right boxplot but the histogram is now gone.
p2 = ggMarginal(data=vdat, x=vname, y=NA, type="boxplot", margins="x")
p2
我如何结合p1和p2的正确部分，使我有p1的直方图和p2的箱线图？
How can I combine the correct parts of p1 and p2 such that I have the histogram from p1 and the boxplot from p2?
我正在尝试类似
p1 + p2
或
ggMarginal(p1, data=vdat, x=vname, y=NA, type="boxplot", margins="x")
但是它们不起作用。
推荐答案
根据 ggMarginal 的文档， p 应该是ggplot 散点图。我们可以在 p 中插入以下行作为第一个geom层：
According to ggMarginal's documentation, p is expected to be a ggplot scatterplot. We can insert the following line as the first geom layer in p:
geom_point(aes(y = 0.01), alpha = 0)
y = 0.01 作为现有图的y轴范围内的值，并且 alpha = 0 确保该图层不可见。
y = 0.01 was chosen as a value within the existing plot's y-axis range, and alpha = 0 ensures this layer isn't visible.
使用此 p 运行代码应该会为您提供带有异常值的箱线图。
Running your code with this p should give you the boxplot with outlier.
p <- ggplot(data=vdat, aes_string(x=vname)) +
geom_point(aes(y = 0.01), alpha = 0) +
geom_histogram(aes(y=stat(density)),
bins=nclass.Sturges(vdat[[vname]])+1,
color="black", fill="steelblue", na.rm=T) +
geom_density(na.rm=T) +
theme_bw()
p1 = ggMarginal(p, type="boxplot", margins = "x")
p1
顺便说一句，在这种情况下，我认为在右侧绘制箱形图确实没有任何意义，因为您尚未为 y 。
By the way, I don't think it really makes sense to plot a boxplot to the right in this instance, since you have not assigned any variable to y.
这篇关于如何在R中使用ggMarginal将箱线图添加到直方图的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！