geom_label_repel()高于geom_line()

Question

因此，我有了这个ggplot，标签是由geom_label_repel()方法定位的。这是相当好的，但我希望标签都在geom_line()之上，而不是在下面，这在某些情况下是这样的。

​

​

这是生成图形的代码：

    ggplot(data,
   aes(x=daydelta,y=day1ret)
   ) + 
  geom_line(color='steelblue',size=2) +
  geom_label_repel(aes(label = paste0(round(day1ret*100,2),"%")),box.padding = unit(0.6, "lines"),label.size = 0.1)+
  scale_x_continuous(limits = c(1,15),breaks=1:15,minor_breaks = 1:15) +
  scale_y_continuous(limits = c(0,1),breaks=seq(0,1,0.1),minor_breaks = seq(0,1,0.1),labels = paste0(seq(0,1,0.1)*100,"%")) +
  ggtitle("RETENTION BY USERID") +
  theme(plot.title = element_text(hjust = 0.5))+
  xlab("Days from Register\n (baseline: Day 1)")+
  ylab("Retention as % of returning users from Day 1")+
  geom_segment(mapping=aes(x=daydelta,y=0,xend=daydelta,yend=day1ret),size=0.5,color="red",linetype=2)

Answer 1

与其用与数据线偏移的标签来乱搞地块，不如将标签放在实际的x值和y值上？这在视觉上要干净得多，不会通过在数据值以外的位置放置标签来分散观众的注意力或误导观众。例如：

library(ggplot2)
library(scales)

dat = data.frame(x=1:10, y=c(1, seq(.3,.1,length=9)))

ggplot(dat, aes(x,y)) +
  geom_line() +
  geom_label(aes(label=paste0(sprintf("%1.1f",y*100),"%")), size=3, label.padding=unit(2,"pt")) +
  scale_y_continuous(labels=percent, limits=c(0,1)) +
  scale_x_continuous(breaks=dat$x) +
  theme_classic()

​

​