如何在R语言ggplot绘图中为双组数据图表添加图例

我用R的ggplot包，基于两个数据集（UserCI为用户组，对应红色；NonUserCI为非用户组，对应蓝色）绘制了带误差棒的散点图，但不知道怎么添加图例。

数据集 #

UserCI（用户组） #

structure(list(Attribute = c("Nscore", "Escore", "Oscore", "Ascore", "Cscore", "Impulsivity", "SS"), Mean = c(0.519519745762712, -0.224147033898305, 0.345051694915254, -0.542761016949153, -0.432290169491526, 0.573723898305084, 0.625454406779661), lower_bound = c(0.345515567755788, -0.421929253136834, 0.173007836159723, -0.743825778750619, -0.620318735695037, 0.417301607369938, 0.461852381381636), upper_bound = c(0.693523923769636, -0.0263648146597761, 0.517095553670785, -0.341696255147686, -0.244261603288014, 0.730146189240231, 0.789056432177685)), class = c("tbl_df", "tbl", "data.frame"), row.names = c(NA, -7L))

NonUserCI（非用户组） #

structure(list(Attribute = c("Nscore", "Escore", "Oscore", "Ascore", "Cscore", "Impulsivity", "SS"), Mean = c(-0.0346437351443125, 0.014794833050368, -0.0236125863044712, 0.0359841765704582, 0.0284564233163561, -0.0306152461799648, -0.0452792359932086), lower_bound = c(-0.0809079532901976, -0.0313902162687121, -0.0700665170072849, -0.00972637107557454, -0.0176800056429342, -0.0748829267378616, -0.089769617527101), upper_bound = c(0.0116204830015727, 0.0609798823694481, 0.0228413443983426, 0.0816947242164909, 0.0745928522756465, 0.0136524343779321, -0.000788854459316299)), class = c("tbl_df", "tbl", "data.frame"), row.names = c(NA, -7L))

当前绘图代码 #

ggplot(UserCI, aes(Attribute, Mean)) + 
    geom_point() + geom_errorbar(aes(ymin = lower_bound, ymax = upper_bound), colour="red") +
    geom_point(data = NonUserCI, aes(Attribute, Mean)) + 
    geom_errorbar(aes(ymin =NonUserCI$lower_bound, ymax = NonUserCI$upper_bound), colour = "blue") 

解决方法 #

ggplot的图例是基于aes映射自动生成的，你当前的代码直接在图层外指定colour参数，无法让ggplot识别分组关系，所以不会生成图例。正确做法是先合并数据集，添加分组标识，再把颜色映射到分组变量。

步骤1：合并数据集 #

给两个数据集分别添加分组列，再合并成一个数据集：

# 给数据集添加分组标签
UserCI$Group <- "用户组"
NonUserCI$Group <- "非用户组"

# 合并两个数据集
combined_data <- rbind(UserCI, NonUserCI)

步骤2：重新绘图并生成图例 #

基于合并后的数据集，把colour放到aes中映射分组变量，ggplot会自动生成图例。还可以通过scale_colour_manual指定你需要的红色和蓝色：

ggplot(combined_data, aes(x = Attribute, y = Mean, colour = Group)) +
    geom_point(size = 2) + # 可调整点的大小
    geom_errorbar(aes(ymin = lower_bound, ymax = upper_bound), width = 0.2) + # 调整误差棒宽度，让图表更美观
    scale_colour_manual(values = c("用户组" = "red", "非用户组" = "blue")) +
    labs(title = "分组属性均值与置信区间", 
         x = "属性", 
         y = "均值", 
         colour = "组别") # 自定义图表各部分的标签

关键逻辑说明 #

之前分开调用图层的方式，相当于手动给两组元素指定颜色，但没有建立“分组-颜色”的映射关系，ggplot无法生成对应的图例。合并数据集后用映射的方式，既符合ggplot的语法逻辑，后续调整分组样式也更便捷。

内容的提问来源于stack exchange，提问作者ano273