如何解决geom_dl()绘制折线图时的标签重叠问题？

下面是几个能有效改善标签重叠问题的具体修改方案：

方案1：调整directlabels的方法参数 #

可以给bumpup添加偏移量，或者改用更智能的布局方法：

• 给bumpup增加x/y方向的偏移，拉开标签间距：
  修改last.bumpup的定义，同时适当扩大y轴上限避免标签超出：

library(ggplot2)
library(directlabels)
library(reshape2) # 补充melt所需包

#Original df
df <- data.frame(names=c('AAAA', 'BBBB', 'CCCC', 'DDDD', 'EEEE'),SP22=c(57, 30, 27, 35, 34),FA22=c(52, 38, 31, 34, 31),SP23=c(49, 32, 30, 29, 31))

#df to long format
longdf <- melt(df, id='names')

last.bumpup <- list("last.points", bumpup(x = 0.2, y = 2))  # x控制水平偏移，y控制垂直偏移

#lineplot
longdf %>%
ggplot() + 
  geom_line(aes(x=variable, y=value, group=names, color=names)) +
  scale_y_continuous(n.breaks=6, limits=c(20, 65))  # 扩大y轴上限
  scale_color_manual(values=c("darkred", "steelblue","black", "coral1", 
                              "darkorchid2")) +
  geom_dl(aes(x=variable, y=value,label=names, color= names),
          method=last.bumpup, cex=0.8) +
  labs(title = "Comparison of...") + 
  xlab(label = "Terms") +
  ylab(label = "Numbers") +
  theme_minimal() +
  theme(legend.position = "none")

• 改用smart.grid布局方法，自动优化标签位置：

longdf %>%
ggplot() + 
  geom_line(aes(x=variable, y=value, group=names, color=names)) +
  scale_y_continuous(n.breaks=6, limits=c(20, 60)) +
  scale_color_manual(values=c("darkred", "steelblue","black", "coral1", 
                              "darkorchid2")) +
  geom_dl(aes(x=variable, y=value,label=names, color= names),
          method="smart.grid", cex=0.8) +  # 替换为smart.grid方法
  labs(title = "Comparison of...") + 
  xlab(label = "Terms") +
  ylab(label = "Numbers") +
  theme_minimal() +
  theme(legend.position = "none")

方案2：使用ggrepel包的智能标签避让 #

ggrepel专门解决标签重叠问题，效果更稳定：
首先安装并加载包：

install.packages("ggrepel")
library(ggrepel)

然后替换geom_dl为geom_text_repel，指定只在最后一个数据点添加标签：

# 过滤出最后一个term的数据，仅给这些点加标签
label_df <- subset(longdf, variable == "SP23")

longdf %>%
ggplot() + 
  geom_line(aes(x=variable, y=value, group=names, color=names)) +
  scale_y_continuous(n.breaks=6, limits=c(20, 65)) +
  scale_color_manual(values=c("darkred", "steelblue","black", "coral1", 
                              "darkorchid2")) +
  geom_text_repel(data = label_df,
                  aes(x=variable, y=value, label=names, color=names),
                  nudge_x = 0.1,  # 标签向右偏移，远离折线
                  force = 2,      # 增强标签避让力度
                  cex=0.8) +
  labs(title = "Comparison of...") + 
  xlab(label = "Terms") +
  ylab(label = "Numbers") +
  theme_minimal() +
  theme(legend.position = "none")

方案3：手动调整标签位置 #

如果需要精准控制，可以给每个系列的标签设置不同的y偏移：

library(dplyr) # 补充case_when所需包

label_df <- subset(longdf, variable == "SP23")
# 手动给每个系列设置y轴偏移量
label_df$y_adjust <- case_when(
  label_df$names == "AAAA" ~ 3,
  label_df$names == "BBBB" ~ 2,
  label_df$names == "CCCC" ~ 1,
  label_df$names == "DDDD" ~ -1,
  label_df$names == "EEEE" ~ -2
)

longdf %>%
ggplot() + 
  geom_line(aes(x=variable, y=value, group=names, color=names)) +
  scale_y_continuous(n.breaks=6, limits=c(20, 65)) +
  scale_color_manual(values=c("darkred", "steelblue","black", "coral1", 
                              "darkorchid2")) +
  geom_text(data = label_df,
            aes(x=variable, y=value + y_adjust, label=names, color=names),
            nudge_x = 0.1,
            cex=0.8) +
  labs(title = "Comparison of...") + 
  xlab(label = "Terms") +
  ylab(label = "Numbers") +
  theme_minimal() +
  theme(legend.position = "none")

内容的提问来源于stack exchange，提问作者RKeithL