根据列名创建多个图表

Question

这是我在 stackoverlow 上的第一个问题，如果我没有遵循正确的问题协议，请纠正我。

我正在尝试为在三个时间点（时间 1、时间 2、时间 3）收集的数据创建一些图表，这些时间点等于 X1...、X2... 和 X3...列名的开头。这些图也由数据框的 $Group 列分隔。

我创建图表没问题，我只有很多变量 (~170)，我想比较时间 1 与时间 2、时间 2 与时间 3 等。所以我想找一个捷径运行这种代码，而不必逐个输入。

如上所述，我创建了像 X1...X2... 这样的变量名，它表示变量被记录的时间，即 X1BCSTCAT = time 1; X2BCSTCAT = 时间 2； X3BCSTCAT = 时间 3。这是我的数据的一小部分示例：

df <- structure(list(ID = structure(1:6, .Label = c("101","102","103","118","119","120"), class = "factor"), 
                   Group = structure(c(1L,1L,1L,2L,2L,2L), .Label = c("C8","TC"), class = "factor"), 
                   Wave = structure(c(1L, 2L, 3L, 4L, 1L, 2L), .Label = c("A","B","C","D"), class = "factor"), 
                   Yr = structure(c(1L, 2L, 1L, 2L, 1L, 2L), .Label = c("3","5"), class = c("ordered", "factor")), 
                   Age.Yr. = c(10.936,10.936, 9.311, 10.881, 10.683, 11.244), 
                   Training..hr. = c(10.667,10.333, 10.667, 10.333, 10.333, 10.333), 
                   X1BCSTCAT = c(-0.156,0.637,-1.133,0.637,2.189,1.229), 
                   X1BCSTCR = c(0.484,0.192, -1.309, 0.912, 1.902, 0.484), 
                   X1BCSTPR = c(-1.773,0.859, 0.859, 0.12, -1.111, 0.12), 
                   X2BCSTCAT = c(1.006, -0.379,-1.902, 0.444, 2.074, 1.006), 
                   X2BCSTCR = c(0.405, -0.457,-1.622, 1.368, 1.981, 0.168), 
                   X2BCSTPR = c(-0.511, -0.036,2.189, -0.036, -0.894, 0.949),
                   X3BCSTCAT = c(1.18, -1.399,-1.399, 1.18, 1.18, 1.18), 
                   X3BCSTCR = c(0.967, -1.622, -1.622,0.967, 0.967, 1.255), 
                   X3BCSTPR = c(-1.282, -1.282, 1.539,1.539, 0.792, 0.792)), 
              row.names = c(1L, 2L, 3L, 4L, 5L,8L), class = "data.frame")

这里是一些工作代码，使用 ggplot 为一个变量的时间 1 与时间 2 数据创建一个图表：

library(ggplot2)

p <- ggplot(df, aes(x=df$X1BCSTCAT, y=df$X2BCSTCAT, shape = df$Group, color = df$Group)) + 
  geom_point() + geom_smooth(method=lm, aes(fill=df$Group), fullrange = TRUE) + 
  labs(title="BCSTCAT", x="Time 1", y = "Time 2") + 
  scale_color_manual(name = "Group",labels = c("C8","TC"),values = c("blue", "red")) +
  scale_shape_manual(name = "Group",labels = c("C8","TC"),values = c(16, 17)) +
  scale_fill_manual(name = "Group",labels = c("C8", "TC"),values = c("light blue", "pink"))

所以我真的在尝试创建某种快捷方式，R 将在其中循环并匹配变量名称 X1... 与 X2... 等等并创建图形。我假设必须有某种方法可以根据匹配的列号进行绘制，例如df[7] vs df[10] 并迭代此过程或通过实际匹配名称进行绘图（其中变量名称的唯一区别是表示时间的数字）。

我以前循环使用 lapply 函数创建单独的图表，但不知道从哪里开始尝试做这个。

Answer 1

使用tidyeval 方法的解决方案。我们需要 ggplot2 v3.0.0（记得重启你的 R 会话）

install.packages("ggplot2", dependencies = TRUE)

首先我们构建一个将列名和组名作为输入的函数。注意 rlang::sym、rlang::quo_name & !!.
然后为 x- 和 y- 值创建 2 个名称向量，以便我们可以使用 purrr::map2 同时循环遍历它们。

library(rlang)
library(tidyverse)

df <- structure(list(ID = structure(1:6, .Label = c("101","102","103","118","119","120"), class = "factor"), 
                   Group = structure(c(1L,1L,1L,2L,2L,2L), .Label = c("C8","TC"), class = "factor"), 
                   Wave = structure(c(1L, 2L, 3L, 4L, 1L, 2L), .Label = c("A","B","C","D"), class = "factor"), 
                   Yr = structure(c(1L, 2L, 1L, 2L, 1L, 2L), .Label = c("3","5"), class = c("ordered", "factor")), 
                   Age.Yr. = c(10.936,10.936, 9.311, 10.881, 10.683, 11.244), 
                   Training..hr. = c(10.667,10.333, 10.667, 10.333, 10.333, 10.333), 
                   X1BCSTCAT = c(-0.156,0.637,-1.133,0.637,2.189,1.229), 
                   X1BCSTCR = c(0.484,0.192, -1.309, 0.912, 1.902, 0.484), 
                   X1BCSTPR = c(-1.773,0.859, 0.859, 0.12, -1.111, 0.12), 
                   X2BCSTCAT = c(1.006, -0.379,-1.902, 0.444, 2.074, 1.006), 
                   X2BCSTCR = c(0.405, -0.457,-1.622, 1.368, 1.981, 0.168), 
                   X2BCSTPR = c(-0.511, -0.036,2.189, -0.036, -0.894, 0.949),
                   X3BCSTCAT = c(1.18, -1.399,-1.399, 1.18, 1.18, 1.18), 
                   X3BCSTCR = c(0.967, -1.622, -1.622,0.967, 0.967, 1.255), 
                   X3BCSTPR = c(-1.282, -1.282, 1.539,1.539, 0.792, 0.792)), 
              row.names = c(1L, 2L, 3L, 4L, 5L,8L), class = "data.frame")

# define a function that accept strings as input
pair_plot <- function(x_var, y_var, group_var) {

  # convert strings to symbols
  x_var <- rlang::sym(x_var)
  y_var <- rlang::sym(y_var)
  group_var <- rlang::sym(group_var)

  # unquote symbols using !! 
  ggplot(df, aes(x = !! x_var, y = !! y_var, shape = !! group_var, color = !! group_var)) + 
    geom_point() + geom_smooth(method = lm, aes(fill = !! group_var), fullrange = TRUE) + 
    labs(title = "BCSTCAT", x = rlang::quo_name(x_var), y = rlang::quo_name(y_var)) +
    scale_color_manual(name = "Group", labels = c("C8", "TC"), values = c("blue", "red")) +
    scale_shape_manual(name = "Group", labels = c("C8", "TC"), values = c(16, 17)) +
    scale_fill_manual(name = "Group",  labels = c("C8", "TC"), values = c("light blue", "pink")) +
    theme_bw()
}

# Test if the new function works
pair_plot("X1BCSTCAT", "X2BCSTCAT", "Group")

# Create 2 parallel lists 
list_x <- colnames(df)[-c(1:6, (ncol(df)-2):(ncol(df)))]
list_x
#> [1] "X1BCSTCAT" "X1BCSTCR"  "X1BCSTPR"  "X2BCSTCAT" "X2BCSTCR"  "X2BCSTPR"

list_y <- lead(colnames(df)[-(1:6)], 3)[1:length(list_x)]
list_y
#> [1] "X2BCSTCAT" "X2BCSTCR"  "X2BCSTPR"  "X3BCSTCAT" "X3BCSTCR"  "X3BCSTPR"

# Loop through 2 lists simultaneously 
# Supply inputs to pair_plot function using purrr::map2
map2(list_x, list_y, ~ pair_plot(.x, .y, "Group"))

示例输出：

#> [[1]]

#> 
#> [[2]]

由 reprex package (v0.2.0) 创建于 2018-05-24。

根据列名创建多个图表

Creating multiple graphs based upon the column names

r

ggplot2

tidyeval