ggrepel标签填充颜色问题

Question

我第一次使用ggplot2，我很难让我用ggrepel创建的标签的颜色动态变化。目前，我的代码如下所示：

    ggplot(tstat) + 
        geom_point(aes(Mu, Sigma),size = 5, color = 'black') + 
        geom_label_repel(aes(Mu, Sigma, label = VarNames, fill = factor(Hemisphere)), fontface = 'bold', color = 'white',
               box.padding = unit(0.25, 'lines'),point.padding = unit(0.5, 'lines')) + 
        geom_rangeframe() + 
        theme_tufte() + 
        xlab(expression(paste(mu, "*"))) + 
        ylab(expression(sigma)) + 
        theme(axis.title.x = element_text(vjust=-0.5), axis.title.y = element_text(vjust=1.5)) + 
        ggtitle("Model Sensitivity by Hemisphere")

一般来说，这是很好的工作，除了我强烈不喜欢它给我的牙膏绿色，因为这两个因素之一。我想规定fill = factor(Hemisphere))线的具体颜色，但我不知道怎么做。

我已经尝试过使用scale_colours_manual函数，但是当我将它包含在第3行的geom_label_repel(.....) paratheses中时，程序抱怨说"ggplot2不知道如何处理ScaleDisclin.Scale/ggproto类的数据“，当我将scale_colours_manual行放置在第3行之外时，它根本没有任何效果，就像在这个例子中，它产生了与上面的图相同的情节：

    ggplot(tstat) + 
        geom_point(aes(Mu, Sigma),size = 5, color = 'black') + 
        scale_colour_manual(values = c('blue', 'red')) + 
        geom_label_repel(aes(Mu, Sigma, label = VarNames, fill = factor(Hemisphere)), fontface = 'bold', color = 'white',
               box.padding = unit(0.25, 'lines'),point.padding = unit(0.5, 'lines')) + 
        geom_rangeframe() + 
        theme_tufte() + 
        xlab(expression(paste(mu, "*"))) + 
        ylab(expression(sigma)) + 
        theme(axis.title.x = element_text(vjust=-0.5), axis.title.y = element_text(vjust=1.5)) + 
        ggtitle("Model Sensitivity by Hemisphere")

我知道一定有办法的，但我不知道。谢谢你的帮助！

编辑:根据要求，我附加了一个tstat的dput()。不是一个大的数据框架。

    structure(list(VarNames = structure(c(4L, 1L, 3L, 2L, 5L, 6L, 
    4L, 1L, 3L, 2L, 5L, 6L), .Label = c("Dry Deposition", "MEGAN Acetone", 
    "MEGAN Terpenes", "Monoterpene Yield", "Ocean", "Photolysis"), class = "factor"), 
    Mu = c(2703.09, 8066.01, 6566.6, 19741.7, 5809.6, 14231.8, 1493.56, 3067.54, 3631.32, 9951.06, 8748.95, 7967.93),
    Sigma = c(3478.28, 8883.23, 7276.49, 18454.4, 6218.8, 14989.7, 1925.14, 3410.27, 4017.64, 9289.57, 9354.64, 8403.1),
    Hemisphere = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L),
    .Label = c("Northern", "Southern"), class = "factor")), 
    .Names = c("VarNames", "Mu", "Sigma", "Hemisphere"), 
    class = "data.frame", row.names = c(NA, -12L))

Answer 1

您可以使用scale_fill_manual：

tstat <- structure(list(VarNames = structure(c(4L, 1L, 3L, 2L, 5L, 6L, 
    4L, 1L, 3L, 2L, 5L, 6L), .Label = c("Dry Deposition", "MEGAN Acetone", 
    "MEGAN Terpenes", "Monoterpene Yield", "Ocean", "Photolysis"), class = "factor"), 
    Mu = c(2703.09, 8066.01, 6566.6, 19741.7, 5809.6, 14231.8, 1493.56, 3067.54, 3631.32, 9951.06, 8748.95, 7967.93),
    Sigma = c(3478.28, 8883.23, 7276.49, 18454.4, 6218.8, 14989.7, 1925.14, 3410.27, 4017.64, 9289.57, 9354.64, 8403.1),
    Hemisphere = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L),
    .Label = c("Northern", "Southern"), class = "factor")), 
    .Names = c("VarNames", "Mu", "Sigma", "Hemisphere"), 
    class = "data.frame", row.names = c(NA, -12L))
library(ggplot2)
library(ggrepel)
library(ggthemes)
ggplot(tstat) + 
  geom_point(aes(Mu, Sigma),size = 5, color = 'black') + 
  geom_label_repel(aes(Mu, Sigma, label = VarNames, fill = factor(Hemisphere)), fontface = 'bold', color = 'white',
         box.padding = unit(0.25, 'lines'),point.padding = unit(0.5, 'lines')) + 
  geom_rangeframe() + 
  theme_tufte() + 
  xlab(expression(paste(mu, "*"))) + 
  ylab(expression(sigma)) + 
  theme(axis.title.x = element_text(vjust=-0.5), axis.title.y = element_text(vjust=1.5)) + 
  ggtitle("Model Sensitivity by Hemisphere") + 
  scale_fill_manual(values = setNames(c("lightblue", "darkgreen"), levels(tstat$Hemisphere)))

​

​