如何调整ggplot的vjust与hjust解决标注偏移/重叠问题

问题根源 #

你之前把所有指标标注都固定在x=Inf, y=-Inf同一个坐标点上，hjust和vjust仅控制文本自身的对齐方式，并非位置偏移参数，所以所有文本挤在同一点，必然出现重叠或偏移出图的情况。

可行解决方案 #

方案1：合并成多行文本（最简洁） #

把所有指标用\n拼接成一个多行字符串，一次性标注在右下角，自动垂直排列，彻底避免重叠：

# 拼接所有指标为多行文本
metrics_text <- paste(
  paste("Model R-squared:", round(best_model_results$model_r2, 2)),
  paste("Test R-squared:", round(best_model_results$test_r2, 2)),
  paste("Correlation:", round(best_model_results$correlation, 2)),
  paste("MSE:", round(best_model_results$MSE, 2)),
  paste("Max Error:", round(best_model_results$max_error, 2)),
  sep = "\n"
)

# 绘制图表并添加标注
gg <- ggplot(pred_results, aes(x = Actual, y = Predicted)) +
  geom_point(color = "blue", alpha = 0.7) +
  geom_abline(intercept = 0, slope = 1, color = "black", linetype = "dashed") +
  labs(title = "Test Set: Actual vs Predicted",
       x = "Actual (%)",
       y = "Predicted (%)") +
  theme_minimal() +
  # 把多行文本放在右下角，hjust=1右对齐，vjust=0让文本从该点向上排列
  annotate("text", x = Inf, y = -Inf, hjust = 1, vjust = 0,
           label = metrics_text, size = 3.5)

方案2：逐个调整标注位置（更灵活） #

如果需要单独控制每个标注的垂直位置，可以基于数据范围设置偏移量，让标注从右下角向上依次排列：

gg <- ggplot(pred_results, aes(x = Actual, y = Predicted)) +
  geom_point(color = "blue", alpha = 0.7) +
  geom_abline(intercept = 0, slope = 1, color = "black", linetype = "dashed") +
  labs(title = "Test Set: Actual vs Predicted",
       x = "Actual (%)",
       y = "Predicted (%)") +
  theme_minimal() +
  # 每个标注y值按数据比例递增，hjust=1右对齐
  annotate("text", x = Inf, y = -Inf, hjust = 1, vjust = 0,
           label = paste("Model R-squared:", round(best_model_results$model_r2, 2))) +
  annotate("text", x = Inf, y = -Inf + diff(range(pred_results$Predicted))*0.03, hjust = 1, vjust = 0,
           label = paste("Test R-squared:", round(best_model_results$test_r2, 2))) +
  annotate("text", x = Inf, y = -Inf + diff(range(pred_results$Predicted))*0.06, hjust = 1, vjust = 0,
           label = paste("Correlation:", round(best_model_results$correlation, 2))) +
  annotate("text", x = Inf, y = -Inf + diff(range(pred_results$Predicted))*0.09, hjust = 1, vjust = 0,
           label = paste("MSE:", round(best_model_results$MSE, 2))) +
  annotate("text", x = Inf, y = -Inf + diff(range(pred_results$Predicted))*0.12, hjust = 1, vjust = 0,
           label = paste("Max Error:", round(best_model_results$max_error, 2)))

这里用diff(range(pred_results$Predicted))*0.03基于数据范围设置偏移，适配不同数据集无需硬编码数值。

方案3：用文本框包裹（更美观） #

如果想让标注更整洁，可借助ggtext包添加带背景的文本框：

library(ggtext)

gg <- ggplot(pred_results, aes(x = Actual, y = Predicted)) +
  geom_point(color = "blue", alpha = 0.7) +
  geom_abline(intercept = 0, slope = 1, color = "black", linetype = "dashed") +
  labs(title = "Test Set: Actual vs Predicted",
       x = "Actual (%)",
       y = "Predicted (%)") +
  theme_minimal() +
  geom_richtext(
    x = Inf, y = -Inf, hjust = 1, vjust = 0,
    label = metrics_text,
    fill = "white", alpha = 0.9, size = 3.5,
    label.padding = unit(0.3, "lines"),
    label.r = unit(0.1, "lines")
  )

内容的提问来源于stack exchange，提问作者NEERAJ YADAV