如何使用 ggtern 包在三元图中的轴上添加标签？答案

【问题标题】：How to add labels on axis in ternary diagram using ggtern package?如何使用 ggtern 包在三元图中的轴上添加标签？
【发布时间】：2017-08-22 13:22:41
【问题描述】：

三元图如下图所示。我想在 R 中使用 ggtern 包添加 Z=60、Z=90 和 Y=60 的标签。

R代码链接是the R code of ternary diagram

【问题讨论】：

我在最新版本中添加了一个新的几何图形，让这种类型的事情变得更容易一些，请查看：geom_text_viewport 和 geom_label_viewport，ggtern.com/2017/07/23/version-2-2-1-released。您仍然需要抑制遮罩，或将其放置在文本图层下方，但是这两个几何图形更直观地用于放置相对于视口的 x 和 y 笛卡尔 [0,1] 限制的标签。

标签： r ggplot2 data-visualization data-analysis ggtern

【解决方案1】：

这不是完美的答案，但我尝试使用annotate 达到您想要的结果，如下所示：

ggtern(data=DATA,aes(x,y,z)) + 
  geom_polygon(aes(fill=Series),alpha=.5,color="black",size=0.25) +
  scale_fill_manual(values=as.character(unique(DATA$Series))) +
  theme(legend.position=c(0,1),legend.justification=c(0,1)) + 
  labs(fill="Region",title="Sample Filled Regions") +
  annotate(geom  = 'text',
           x     = c(0.1, 1/3, 0.0),
           y     = c(0.0, 0.0, 1.5),
           z     = c(0.5, 1/3, 1.0),
           angle = c(0, 0, 0),
           vjust = c(2.5, 2.5, -1.5),
           hjust = c(0.0, -0.2, 0.0),
           label = c("Z=90","Z=60","Y=60"),
           color = c("black","gray",'orange')) + # for inspection
  theme_nomask()   # allows drawing beyond the borders

这会产生以下图片：

【讨论】：

【解决方案2】：

一个相当“原始”的解决方案是使用 grobs。
在找到包含 x、y 和 z 标签的文本 grobs（最初放置在三角形的顶点）后，我们将每个标签移动到所需的位置。

library(ggtern)
library(grid)
g <- data.frame(y=c(1,0,0),
                x=c(0,1,.4),
                z=c(0,0,.6),         Series="Green")
p <- data.frame(y=c(1,0.475,0.6),
                x=c(0,0.210,0),
                z=c(0,0.315,.4),         Series="Red")
q <- data.frame(y=c(0.575,0.475,0.0,0.0),
                x=c(0.040,0.210,0.4,0.1),
                z=c(0.385,0.315,0.6,0.9),         Series="Yellow")
f <- data.frame(y=c(0.6,0.575,0.0,0.0),
                x=c(0.0,0.040,0.1,0.0),
                z=c(0.4,0.385,0.9,1.0),         Series="Blue")
DATA = rbind(g, p, q, f)

p <- ggtern(data=DATA,aes(x,y,z)) + 
  geom_polygon(aes(fill=Series),alpha=.5,color="black",size=0.25) +
  scale_fill_manual(values=as.character(unique(DATA$Series))) +
  theme(legend.position=c(0,1),legend.justification=c(0,1)) + 
  labs(fill="Region",title="Sample Filled Regions")+
  xlab("X=60")+ylab("Y=60")+zlab("Z=90")

gt <- ggplot_gtable(ggplot_build(p))
grobPanel <- gt$grobs[[which(gt$layout$name == "panel")]] 
grobLabels <- grobPanel$children[[length(grobPanel$children)]]

# X-axes label
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[2]]$x <- unit(0.6,"npc")
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[2]]$y <- unit(0.1,"npc")
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[2]]$gp$fontsize <- 14

# Y-axes label
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[1]]$x <- unit(0.7,"npc")
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[1]]$y <- unit(0.6,"npc")
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[1]]$gp$fontsize <- 14

# Z-axes label
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[3]]$x <- unit(0.75,"npc")
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[3]]$y <- unit(0.1,"npc")
gt$grobs[[which(gt$layout$name == "panel")]]$children[[length(grobPanel$children)]]$children[[3]]$gp$fontsize <- 14

grid.draw(gt)

【讨论】：