如何用cowplot创建带侧边/顶部标题及通用图例的Choropleth面板

要实现带侧边标题、顶部标题和通用图例的2x2 Choropleth面板，可按以下步骤调整代码：

1. 调整图例生成函数，返回grob对象 #

tmap生成的图例需要转为grid图形对象（grob）才能与cowplot的布局函数兼容，修改图例函数：

create_legend <- function(data) {
  legend <- tm_shape(texas) +
    tm_borders() +
    tm_shape(CPC_grids_deviations) +
    tm_fill(col = data, palette = "Blues", title = "Count", style = "fixed", breaks = c(0, 1, 21, 41, 61, 81, 101, 121, 140)) +
    tm_layout(legend.only = TRUE, legend.position = c(0.5, 0.5))
  return(tmap_grob(legend))
}

2. 生成所有图形组件 #

先生成四个地图grob和图例grob：

# 生成四个子地图
p1 <- create_grids_deviation_map("total_cpc_90")
p2 <- create_grids_deviation_map("total_chirps_90")
p3 <- create_grids_deviation_map("total_cpc_70")
p4 <- create_grids_deviation_map("total_chirps_70")

# 生成统一图例
legend_grob <- create_legend("total_cpc_90")

3. 构建带标题的布局 #

通过ggdraw和plot_grid逐步组合布局，添加列标题、侧边标题和整体标题：

library(cowplot)

# 给顶部子图添加列标题（CPC/CHIRPS）
p1_labeled <- ggdraw() + draw_plot(p1) + draw_text("CPC", x = 0.5, y = 1.02, size = 12, fontface = "bold")
p2_labeled <- ggdraw() + draw_plot(p2) + draw_text("CHIRPS", x = 0.5, y = 1.02, size = 12, fontface = "bold")

# 组合第一行，并添加侧边标题"90"
row1 <- plot_grid(p1_labeled, p2_labeled, ncol = 2)
row1_with_side <- ggdraw() + draw_plot(row1) + draw_text("90", x = -0.05, y = 0.5, angle = 90, size = 12, fontface = "bold")

# 组合第二行，并添加侧边标题"70"
row2 <- plot_grid(p3, p4, ncol = 2)
row2_with_side <- ggdraw() + draw_plot(row2) + draw_text("70", x = -0.05, y = 0.5, angle = 90, size = 12, fontface = "bold")

# 组合两行成主面板，添加整体顶部标题
main_panel <- plot_grid(row1_with_side, row2_with_side, nrow = 2, rel_heights = c(1, 1))
main_panel_with_title <- ggdraw(main_panel) + draw_text("Texas Deviation Counts", x = 0.5, y = 1.03, size = 14, fontface = "bold")

# 最后整合主面板和图例，调整宽度比例
final_plot <- plot_grid(main_panel_with_title, legend_grob, ncol = 2, rel_widths = c(4, 1))

# 输出最终图形
print(final_plot)

关键说明 #

• 用tmap_grob()转换tmap对象为grob，确保能被cowplot识别
• 通过ggdraw()的draw_text()添加各类标题，angle = 90实现垂直侧边标题
• 调整rel_widths和rel_heights参数控制各区域的尺寸比例，让布局更协调

内容的提问来源于stack exchange，提问作者Ramaraja