在单个图中为一个连续的 Y 变量绘制多个 X 变量

Question

我正在尝试绘制一个线图，在 x-axis 中有两个 x 变量，在 y-axis 中有一个连续的 y 变量。 x1 和 x2 的计数不同。 df 如下所示-

df <- structure(list(val = c(3817,2428,6160,6729,7151,7451,6272,7146,7063,6344,5465,6169,7315,6888,7167,6759,4903,6461,7010,7018,6920,3644,6541,31862,31186,28090,28488,29349,28284,25815,23529,20097,19945,22118), type = c("1wt", "1wt", "3wt", "3wt", "3wt", "5wt", "5wt", "7wt", "7wt", "7wt","10wt","10wt","10wt","15wt","15wt","20wt","20wt","25wt","25wt","25wt","30wt","30wt","30wt","20m","20m","15m","15m","15m","10m","10m","5m", "5m", "5m", "5m"), group = c("A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B")), row.names = c(NA, 34L), class = "data.frame")

其中 x 个变量是-

x1 <- factor(df$type, levels = c("1wt", "3wt", "5wt", "7wt", "10wt", "15wt", "20wt", "25wt", "30wt")) 和

x2 <- factor(df$type, levels = c("20m", "15m","10m","5m"))

我想根据 x 轴上的 df$group 和 df$val 上的不同颜色和图例为 x1 和 x2 设置单独的行y 轴。你能帮我做这个吗？提前致谢。

Answer 1

编辑：在下方添加

这是一种假设意图是将 A 组中可能类型值的范围与 B 组中可能类型值的范围进行映射的方法。

可以手动添加标签，但我认为没有任何简单的方法可以在一个图中同时使用两个分类 x 轴。

df2 <- df %>%
  mutate(x = case_when(type == "1wt" ~ 0,
                       type == "3wt" ~ 1,
                       type == "5wt" ~ 2,
                       type == "7wt" ~ 3,
                       type == "10wt" ~ 4,
                       type == "15wt" ~ 5,
                       type == "20wt" ~ 6,
                       type == "25wt" ~ 7,
                       type == "30wt" ~ 8,

                       type == "20m"  ~ 0/3 * 8,
                       type == "15m"  ~ 1/3 * 8,
                       type == "10m"  ~ 2/3 * 8,
                       type == "5m"   ~ 3/3 * 8))


ggplot(df2, aes(x, val, color = group, group = group)) + 
  geom_point() +
  geom_smooth(method = lm)

第二种方法

听起来 OP 想以某种方式在数字上使用 type 值。如果它们没有按照所描述的方式相互联系，我怀疑将它们绘制成好像它们是一样的会产生误导。（请参阅 here 了解为什么会出现这种情况的讨论。）

就是说，这就是您可以做到的。首先，这是一种仅按原样使用 type 的数字部分的方法。请注意，与 B 组相关联的 "m" 在底部，而 "wt" 在顶部，与 A 组相关联，如以下 OP 注释中添加的示例所示。我在坐标轴上添加了颜色以阐明这一点。这在视觉上有点违反直觉，因为与顶轴相关的点在底部，反之亦然。

df2 <- df %>%
  # First, let's take the number used in "type" without adjustment
  mutate(x_unadj = parse_number(type))

ggplot(df2, aes(x_unadj, val, color = group, group = group)) + 
  geom_point() +
  geom_smooth(method = lm) + # Feel free to use other smoothing method, but
                             # not obvious to me what would be improvement.
  scale_x_continuous("m", sec.axis = sec_axis(~., name = "wt")) +
  theme(axis.text.x.bottom  = element_text(color = "#00BFC4"),
        axis.title.x.bottom = element_text(color = "#00BFC4"),
        axis.text.x.top     = element_text(color = "#F8766D"),
        axis.title.x.top    = element_text(color = "#F8766D"))

如果这不令人满意，我们可以使用

颠倒两个轴的顺序

scale_x_reverse("m", sec.axis = sec_axis(~., name = "wt")) +

使用 ggplot 3.1.0（自 2018 年 10 月起），我无法让辅助 x 轴向与主轴相反的方向移动。 from 2017 doesn't seem to work with this version any more. As of Dec 2018, there is a proposed fix 正在审查旨在解决此问题。

在单个图中为一个连续的 Y 变量绘制多个 X 变量

Plotting multiple X variables for one continuous Y variable in a single plot

plot

r

ggplot2

melt

rbind