代码之家 › 专栏 › 技术社区 › blazej

在4个因素的组合上迭代3路交互ggplot调用

ggplot2 r

blazej · 技术社区 · 6 年前

免责声明:更好的标题建议非常受欢迎

我很难搞清楚我该如何迭代 ggplot2 调用3个因素的不同组合。我可以根据数据来描述这个。。。 但如果有不清楚的地方,请询问/评论(英语不是我的主要语言)

给定这样的数据帧:

library(ggplot2)
library(dplyr)
library(gtools)

# Make up data
set.seed(1)
df <- data.frame(y = rnorm(16),
                 factor1 = gl(2, 1, 16), 
                 factor2 = gl(2, 2, 16),
                 factor3 = gl(2, 4, 16),
                 factor4 = gl(2, 8, 16))

考虑到4个因素( factor1:factor4 )由3个元素组成的24个独特组合,我称之为:

x axis
lines
facets

储存在 p :

p <- data.frame(
      permutations(4,3, c("factor1", "factor2", "factor3", "factor4")))
names<- c("x axis", "lines", "facets")
colnames(p) <- names

> head(p)
   x axis   lines  facets
1 factor1 factor2 factor3
2 factor1 factor2 factor4
3 factor1 factor3 factor2
4 factor1 factor3 factor4
5 factor1 factor4 factor2
6 factor1 factor4 factor3

如何自动绘制24个图形,以便:

价值观 X轴 , 线 和 面 从每一排 第页 逐一地
df$y 总结( mean(y, na.rm=T )跨越相应因素
显示了一个图

如果我在哪里手动绘制,则第四个图(24个图中的一个)将如下所示:

p[4,]
   x axis   lines  facets
4 factor1 factor3 factor4

df %>% group_by(factor1, factor3, factor4) %>%
      summarise(y=mean(y, na.rm=T)) %>% 
      ggplot(aes(y=y, x = factor1)) +
      geom_line(aes(group=factor3, color=factor3)) +
      facet_grid(~factor4)

而第24天是这样的:

> p[24,]
    x axis   lines  facets
24 factor4 factor3 factor2

    df %>% group_by(factor4, factor3, factor2) %>%
          summarise(y=mean(y, na.rm=T)) %>% 
          ggplot(aes(y=y, x = factor4)) +
          geom_line(aes(group=factor3, color=factor3)) +
          facet_grid(~factor2)

有人能帮我解决这个问题吗?

编辑1

我已经能写 for() 循环以聚合数据,但对于resons,我并没有低估所有24个数据帧都是相同的:

 for(i in 1:nrow(p)){
      xaxis <- as.character(p[i,1])
      lines <- as.character(p[i,2])
      facets <- as.character(p[i,3])
      colnames(df)[colnames(df) == xaxis[i]] <- "xaxis"
      colnames(df)[colnames(df) == lines[i]] <- "lines"
      colnames(df)[colnames(df) == facets[i]] <- "facets"
      print(df %>% group_by(xaxis, lines, facets) %>%
                  summarise(y=mean(y, na.rm=T)))
      }

2 回复 | 直到 6 年前

Parfait 6 年前

考虑R基 Map (包装到 mapply )作为通过置换数据帧按元素迭代的替代方法。

p <- data.frame(permutations(4,3, c("factor1", "factor2", "factor3", "factor4")),
                stringsAsFactors = FALSE) %>% 
         setNames(c("xaxis", "lines", "facets"))

runplots <- function(x, l, f) {      
  df2 <- df[c("y", x, l, f)] %>% 
           setNames(c("y", "xaxis", "lines", "facets"))

  print(df2 %>% 
          group_by(xaxis, lines, facets) %>%
          summarise(y=mean(y, na.rm=T)) %>%
          ggplot(aes(y=y, x = xaxis)) +
            geom_line(aes(group=lines, color=lines)) +
            facet_grid(~facets))
}

plot_list <- Map(runplots, p$xaxis, p$lines, p$facets)

因为你保留了一个输出为 地图 ,您可以使用该对象在其他地方继续使用。均匀移动 print 调用图如下所示:

# BY INDEX
plot_list[[1]]
plot_list[[2]]
...

# BY NAME (AFTER RENAMING LIST)
names(plot_list) <- gsub("factor", "f", paste(p$xaxis, p$lines, p$facets, sep="_"))
plot_list$f1_f2_f3
plot_list$f1_f2_f4
...

blazej 6 年前

好啊。。。我花了半天的时间反复试验,但最终还是成功了

组成数据

library(ggplot2)
library(dplyr)
library(gtools)


set.seed(1)
df <- data.frame(y = rnorm(16),
                 factor1 = gl(2, 1, 16), 
                 factor2 = gl(2, 2, 16),
                 factor3 = gl(2, 4, 16),
                 factor4 = gl(2, 8, 16))

p <- data.frame(
      permutations(4,3, c("factor1", "factor2", "factor3", "factor4")))
names<- c("xaxis", "lines", "facets")
colnames(p) <- names

绘制绘图

for(i in 1:nrow(p)){
      xaxis <- as.character(p[i,1])
      lines <- as.character(p[i,2])
      facets <- as.character(p[i,3])
      df2 <- df %>% rename(xaxis= colnames(df)[colnames(df) == xaxis],
                         lines= colnames(df)[colnames(df) == lines],
                         facets= colnames(df)[colnames(df) == facets])
      print(df2 %>% 
                  group_by(xaxis, lines, facets) %>%
                  summarise(y=mean(y, na.rm=T)) %>%
                  ggplot(aes(y=y, x = xaxis)) +
                  geom_line(aes(group=lines, color=lines)) +
                  facet_grid(~facets))
}

我会把这个作为一个开放的答案,如果有人想给我展示更优雅的东西(只是为了我的好奇心),我非常乐意接受。