指定多个时从ggplot颜色图例中删除一个因子

时间：2014-02-11 14:51:54

标签： r ggplot2

我将两个因素映射到ggplot中的color参数，并希望从图例中删除一个。

我有这些数据。

Data <- structure(list(StudyArea = c("AAA", "BBB", "CCC", "AAA", "BBB", 
"CCC", "AAA", "BBB", "CCC"), Obs = c(190L, 481L, 219L, 190L, 
481L, 219L, 190L, 481L, 219L), InSituPred = c(180, 462, 199, 
180, 462, 199, 180, 462, 199), InSituSE = c(9.57382456553708, 
16.5306359391421, 9.51070020039693, 9.57382456553708, 16.5306359391421, 
9.51070020039693, 9.57382456553708, 16.5306359391421, 9.51070020039693
), variable = c("ExSituAAA", "ExSituAAA", "ExSituAAA", "ExSituBBB", 
"ExSituBBB", "ExSituBBB", "ExSituCCC", "ExSituCCC", "ExSituCCC"
), value = c(NA, 500, 172, 156, NA, 161, 200, 562, NA), SE = c(20.8552264204181, 
16.1382421185167, 21.43693858142, 20.8552264204181, 16.1382421185167, 
21.43693858142, 20.8552264204181, 16.1382421185167, 21.43693858142
)), .Names = c("StudyArea", "Obs", "InSituPred", "InSituSE", 
"variable", "value", "SE"), row.names = c(NA, -9L), class = "data.frame")

使用下面的代码，可以制作如下图。

ggplot(Data)+
  geom_point(aes(x=StudyArea, y=value, color=variable),size=3, shape=1)+
  geom_errorbar(aes(x=StudyArea, ymin=value-SE, ymax=value+SE, color=variable),lty = 2, cex=0.75)+
  geom_point(aes(x=StudyArea, y=InSituPred, color=StudyArea),size=3, shape=1)+
  geom_errorbar(aes(x=StudyArea, ymin=InSituPred-InSituSE, ymax=InSituPred+InSituSE, color=StudyArea),lty=1,cex=0.75)+
  geom_point(aes(x=StudyArea, y=Obs, color=StudyArea),shape="*",size=12)

Fig

我想从图例（AAA：CCC）中删除StudyArea颜色，以便它只包含ExSituAAA，ExSituBBB，ExSituCCC。

ADDITION 使用下面的代码（以及@shadow的有用评论）我可以创建下图。

p <- ggplot(Data, aes(x=StudyArea))+
  geom_point(aes(y=value, color=variable),size=3, shape=1)+
  geom_errorbar(aes(ymin=value-SE, ymax=value+SE, color=variable),lty = 2, cex=0.75)+
  geom_point(aes(y=InSituPred, color=StudyArea),size=3, shape=1)+
  geom_errorbar(aes(ymin=InSituPred-InSituSE, ymax=InSituPred+InSituSE, color=StudyArea),lty=1,cex=0.75)+
  geom_point(aes(y=Obs, color=StudyArea),shape="*",size=12) +
  scale_color_discrete(breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC"))

p +  scale_color_manual(name="Study Area \nPrediction", 
                        values=c("red", "blue", "darkgreen","red","blue","darkgreen"), 
                        breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC"))

FigureII

我想在图例中添加linetype，指定实线= InSitu和虚线= ExSitu。

我使用linetype手动指定了aes（而不是lty中的因素），因为我还需要指定color。查看下面的head，数据$ SE为lty = 2，数据$ InSituSE为lty 1.

＆GT; 

 head(Data)
  StudyArea Obs InSituPred  InSituSE  variable value       SE
1       AAA 190        180  9.573825 ExSituAAA    NA 20.85523
2       BBB 481        462 16.530636 ExSituAAA   500 16.13824
3       CCC 219        199  9.510700 ExSituAAA   172 21.43694
4       AAA 190        180  9.573825 ExSituBBB   156 20.85523
5       BBB 481        462 16.530636 ExSituBBB    NA 16.13824
6       CCC 219        199  9.510700 ExSituBBB   161 21.43694

因此，有关如何使用实线添加linetype图例的任何建议= InSitu和虚线= ExSitu将不胜感激。

示例：我想添加...

figIII

提前致谢。

1 个答案:

答案 0 :(得分：2)

您可以使用?scale_color_discrete指定中断。在您的情况下，这可能类似于以下内容：

ggplot(Data, aes(x=StudyArea))+
  geom_point(aes(y=value, color=variable),size=3, shape=1)+
  geom_errorbar(aes(ymin=value-SE, ymax=value+SE, color=variable),lty = 2, cex=0.75)+
  geom_point(aes(y=InSituPred, color=StudyArea),size=3, shape=1)+
  geom_errorbar(aes(ymin=InSituPred-InSituSE, ymax=InSituPred+InSituSE, color=StudyArea),lty=1,cex=0.75)+
  geom_point(aes(y=Obs, color=StudyArea),shape="*",size=12) +
  scale_color_discrete(breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC"))

编辑：是的，可以指定颜色。因为我真的不明白你想要什么着色方案，所以这里有一些例子（并非所有都是完全认真的）。 

p <- ggplot(Data, aes(x=StudyArea))+
  geom_point(aes(y=value, color=variable),size=3, shape=1)+
  geom_errorbar(aes(ymin=value-SE, ymax=value+SE, color=variable),lty = 2, cex=0.75)+
  geom_point(aes(y=InSituPred, color=StudyArea),size=3, shape=1)+
  geom_errorbar(aes(ymin=InSituPred-InSituSE, ymax=InSituPred+InSituSE, color=StudyArea),lty=1,cex=0.75)+
  geom_point(aes(y=Obs, color=StudyArea),shape="*",size=12) 
p +  scale_color_manual(name="Study Area \nPrediction", 
                        values=c("red", "blue", "darkgreen","red","blue","darkgreen"), 
                        breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC"))
p +  scale_color_manual(name="Study Area \nPrediction", 
                        values=c("black", "black", "black", "red","blue","darkgreen"), 
                        breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC"))
p +  scale_color_manual(name="Study Area \nPrediction", 
                        values=c("white", "yellow", "pink", "red","blue","darkgreen"), 
                        breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC"))

附加：如果您在绘图前清理和重组数据，这将更容易。这是我的豁免：

df <- with(Data, data.frame(area=rep(StudyArea, 2),
                            exarea=c(variable,rep(variable[c(1,4,7)], 3)),
                            value=c(value, InSituPred), 
                            se=c(SE, InSituSE), 
                            obs = rep(Obs, 2),
                            situ=rep(c("in", "ex"), each=nrow(Data))))
df <- df[!duplicated(df),]

然后绘图变得更容易：

p <- ggplot(df, aes(x=area))+
  geom_point(aes(y=value, color=exarea),size=3, shape=1)+
  geom_errorbar(aes(ymin=value-se, ymax=value+se, color=exarea, lty=situ), cex=0.75)+                   
  geom_point(aes(y=obs, color=exarea),shape="*",size=12) 

p +  scale_color_manual(name="Study Area \nPrediction", 
                    values=c("red", "blue", "darkgreen"), 
                    breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC")) +
   scale_linetype_manual(name="Situ", 
                         values=c(1,2), 
                         breaks=c("in", "ex"), 
                         labels=c("InSitu", "ExSitu"))

EDIT2：可以使用原始数据。您必须将lty放在aes - 函数中，然后像以前一样使用scale_linetype_manual。在这里：

p <- ggplot(Data, aes(x=StudyArea))+
  geom_point(aes(y=value, color=variable),size=3, shape=1)+
  geom_errorbar(aes(ymin=value-SE, ymax=value+SE, color=variable, lty="2"), cex=0.75)+
  geom_point(aes(y=InSituPred, color=StudyArea),size=3, shape=1)+
  geom_errorbar(aes(ymin=InSituPred-InSituSE, ymax=InSituPred+InSituSE, color=StudyArea, lty="1"),cex=0.75)+
  geom_point(aes(y=Obs, color=StudyArea),shape="*",size=12) 
p +  scale_color_manual(name="Study Area \nPrediction", 
                        values=c("red", "blue", "darkgreen","red","blue","darkgreen"), 
                        breaks=c("ExSituAAA", "ExSituBBB", "ExSituCCC")) + 
  scale_linetype_manual(name="Situ", 
                        values=c(1,2), 
                        breaks=c("1", "2"), 
                        labels=c("InSitu", "ExSitu"))

通常更好的做法是重组数据。如果您想对此代码进行任何更改，那将很难做到。代码已经很难阅读了。因此，如果您可以重新构建数据集（通常是这样），那么请考虑采用上述方法。

相关问题
最新问题