前言

面积图是在折线图的基础上形成的，它将线段与坐标轴之间的区域用颜色或条纹来填充，以凸显数据的变换趋势。

以 x 轴作为绘制方向，可以使用 geom_ribbon() 函数来绘制一个指定 y 轴区域（即设置 ymin 和 ymax 参数）的图形

使用 geom_area() 绘制线条与 x 轴之间的范围（即 ymin=0, ymax=y）

如果以 y 轴作为绘制方向，那么相应的范围参数则变为 xmin 和 xmax

示例

单数据面积图

df <- tibble(x = sample(30:50, 20), y = sample(1:20, 20))

p1 <- ggplot(df, aes(x, y)) +
  geom_line() +
  geom_area(fill = 'blue', alpha = 0.5)

p2 <- ggplot(df, aes(x, y)) +
  geom_line() +
  geom_ribbon(aes(ymin = y - 1, ymax = y + 1), fill = 'lightgreen', alpha = 0.5)

p3 <- ggplot(df, aes(x, y)) +
  geom_line(orientation = 'y') +
  geom_area(fill = 'blue', alpha = 0.5, orientation = 'y')

p4 <- ggplot(df, aes(x, y)) +
  geom_line(orientation = 'y') +
  geom_ribbon(aes(xmin = x - 1, xmax = x + 1), orientation = 'y',
              fill = 'lightgreen', alpha = 0.5)

plot_grid(p1, p2, p3, p4)

多数据堆积和百分比面积图

p <- subset(economics_long, variable %in% c("pce", "unemploy")) %>%
  ggplot(aes(x = date))

p1 <- p + geom_area(aes(y = value01, fill = variable), alpha = 0.4, position = 'stack')

p2 <- p + geom_area(aes(y = value01, fill = variable), alpha = 0.4, position = 'fill')

plot_grid(p1, p2)

绘制两条曲线之间的区域面积

p <- subset(economics_long, variable %in% c("pce", "unemploy")) %>%
  select(c(date, variable, value01)) %>%
  pivot_wider(names_from = variable, values_from = value01) %>%
  ggplot(aes(x = date)) + 
  geom_ribbon(aes(ymin = if_else(pce > unemploy, unemploy, pce), 
                    ymax = if_else(pce > unemploy, pce, unemploy)),
                fill = "#decbe4", colour = "black")
p

但有时候，我们可能想要区分两条曲线大小关系，体现在图形上就是，曲线 1 大于曲线二的区域设置为一种颜色，曲线 1 小于曲线 2 的区域设置为另一种不同的颜色

p + geom_ribbon(aes(ymin = pmin(unemploy, pce), 
                    ymax = pmax(pce, unemploy),
                    fill = pce > unemploy,
                    colour = pce > unemploy),
                show.legend = FALSE, na.rm = TRUE) +
  scale_fill_manual(values = c("#8dd3c7", "#fdb462"))

但是，从图上我们看到，不同的区域之间会有粘连，那怎么去除这些粘连呢？

我的解决办法是，将两种区域分开绘制，在绘制某一种区域时，将相反区域的值都设置为 NA

df <- subset(economics_long, variable %in% c("pce", "unemploy")) %>%
  select(c(date, variable, value01)) %>%
  pivot_wider(names_from = variable, values_from = value01) %>%
  mutate(low_min = pmin(pce, unemploy), low_max = pmax(pce, unemploy),
         high_min = low_min, high_max = low_max)

df$low_min[df$pce > df$unemploy] = NA
df$low_max[df$pce > df$unemploy] = NA

df$high_min[df$pce <= df$unemploy] = NA
df$high_max[df$pce <= df$unemploy] = NA

ggplot(df, aes(x = date)) +
  geom_ribbon(aes(ymin = low_min, ymax = low_max), 
              fill = "#8dd3c7", alpha = 0.7) +
  geom_ribbon(aes(ymin = high_min, ymax = high_max), 
              fill = "#fdb462", alpha = 0.7) +
  geom_line(aes(y = pce), colour = "#fb8072", size = .75) + 
  geom_line(aes(y = unemploy), colour = "#80b1d3", size = .75)