如何在ggplot2图表的底部插入自定义图例

Question

我正在创建一个带有模型，数据点和几条水平线的ggplot2图。我需要能够在图表的底部放置一个图例，表示蓝色线是模型，红色是最大值，橙色是阈值等。

到目前为止，我有这个：

ggplot(df, aes(DATE, RELATIVE_PERCENT)) + 
       geom_point(colour="black", size=0.8) + 
       geom_smooth(method="loess", size=1, colour="blue", se=T) + 
       scale_x_datetime(breaks = "1 hours") + 
       theme_bw() + 
       geom_hline(aes(yintercept = 100), col="red", size=1.5) + 
       ylab("Relative CPU Utilization") + 
       xlab("Date") + 
       ggtitle("DB CPU Utilization") + 
       geom_hline(aes(yintercept = 70), col="orange", size=1.5)

这是我的样本数据框

dput（DF）

structure(list(DATE = structure(c(1365717060, 1365717090, 1365717120, 
1365717150, 1365717180, 1365717210, 1365717240, 1365717270, 1365717300, 
1365717330, 1365717360, 1365717390, 1365717420, 1365717450, 1365717480, 
1365717510, 1365717540, 1365717570, 1365717600, 1365717630), class = c("POSIXct", 
"POSIXt"), tzone = ""), RELATIVE_PERCENT = c(26, 26, 26, 26, 
26, 26, 25, 26, 26, 27, 25, 26, 26, 26, 26, 26, 26, 27, 27, 27
)), .Names = c("DATE", "RELATIVE_PERCENT"), class = "data.frame", row.names = 5740:5759)

Answer 1

我可能接近这样的传奇：

df <- data.frame(x = 1:10,
                 y = runif(10))

max_min <- data.frame(y = c(0,1),
                      grp = c('min','max'))

ggplot(df)+geom_smooth(aes(x = x,y = y,colour = "model")) + 
    geom_hline(data = max_min,aes(yintercept = y,colour = grp))

Answer 2

我不会把它称为优雅，但这是一个有效的解决方案：

# generate general plot
p <- ggplot(df, aes(DATE, RELATIVE_PERCENT)) +

# add the points and the smoothing
geom_point(aes(colour = "data"), size=0.8) +
geom_smooth(aes(colour = "smoothing"), method="loess", size=1, se=T, show_guide=TRUE) +

# add the lines
geom_hline(aes(yintercept = 100, colour ='overkill'), show_guide = TRUE, size=1.5) +
geom_hline(aes(yintercept = 70, colour ='warning'), show_guide = TRUE, size=1.5) +

# adjust the colours to those you wanted
scale_colour_manual(values = c("black","red", "blue", "orange"))

# stick the legend on the bottom
+ theme( legend.position = "bottom")

现在，有关其中任何一项的详细信息，请查看ggplot2 documentation。如果你想玩传说，R Cookbook section on legends可能会有所帮助。