在ggplot中混合线和散点图

Question

我环顾四周,但我对这个感到难过.我找不到绘制与散点图无关的线的方法.以下是我的一些数据和代码,以澄清问题.我有以下表格的数据

> head(allData)
  AnnounceDate MarketProbability DealStatus binary BrierScore
1   2000-04-10         0.3333333   Complete      1  0.2340565
2   2000-06-14         0.2142857   Complete      1  0.3618200
3   2000-06-26         0.6846154   Complete      1  0.3690167
4   2000-06-16         0.1875000   Complete      1  0.4364041
5   2000-10-05         0.9555556   Complete      1  0.3078432
6   2000-10-19         0.8500000   Complete      1  0.2670799

我想创建一个MarketProbabilities与the 的散点图AnnounceDate,并确定DealStatus是Completed或是Terminated使用颜色.

(ggplot(data=allData, aes(x=AnnounceDate, y=MarketProbability, colour=DealStatus))
 + geom_point() + scale_colour_hue(h = c(180,0)))

我也想覆盖滚动的Brier得分,我试过了

(ggplot(data=allData, aes(x=AnnounceDate, y=MarketProbability, colour=DealStatus))
 + geom_point() + scale_colour_hue(h=c(180,0))
 + geom_line(aes(x=AnnounceDate, y=BrierScore)))

我很困惑为什么要绘制两条彩色线条？如何将该线与先前绘制的数据分离？

Answer 1

正如@MattBagg指出的那样,通过将aes()for颜色移出主ggplot()呼叫并进入geom_point()呼叫来解决此问题.

library(ggplot2)

allData = read.table(header=TRUE,
colClasses=c("Date", "numeric", "character", "numeric", "numeric"),
text="AnnounceDate MarketProbability DealStatus binary BrierScore
2000-04-10         0.3333333   Complete      1  0.2340565
2000-06-14         0.2142857   Complete      1  0.3618200
2000-06-26         0.6846154   Complete      1  0.3690167
2000-06-16         0.1875000   Complete      1  0.4364041
2000-10-05         0.9555556   Complete      1  0.3078432
2000-10-19         0.8500000   Complete      1  0.2670799")

p1 = ggplot(data=allData, aes(x=AnnounceDate)) +
     geom_point(aes(y=MarketProbability, colour=DealStatus)) + 
     scale_colour_hue(h = c(180,0)) + 
     geom_line(aes(y=BrierScore))

ggsave(filename="plot_1.png", plot=p1, height=2.5, width=5)

Answer 2

ggplot()中的aes()由后续的geom继承.你指定了color = DealStatus,它由geom_line()继承,它需要制作两行来制作两种颜色.我无法测试这个,因为我在手机上,但尝试将颜色分配从主ggplot移动到geom_point():

ggplot(data = allData, aes(x=AnnounceDate)) +
     geom_point(aes(colour=DealStatus, y=MarketProbability)) + 
     scale_colour_hue(h = c(180,0)) +
     geom_line(aes(y=BrierScore, group=1))

我正在添加一个常量作为组参数来告诉它连接所有点.否则,ggplot有时会做出错误的猜测.