在ggplot中的每个构面内手动订购x轴标签 [英] Manually order x-axis labels within each facet in ggplot

问题描述

我正在尝试手动重新排列每个构面内的x轴标签.

I am trying to manually reorder the x-axis labels within each facet.

数据如下:

df = structure(list(block = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 
4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 5L, 5L, 5L), .Label = c("1", 
"2", "3", "4", "5"), class = "factor"), item = structure(c(14L, 
15L, 28L, 29L, 30L, 31L, 32L, 15L, 16L, 17L, 18L, 19L, 20L, 21L, 
15L, 22L, 23L, 24L, 25L, 26L, 27L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 
1L, 8L, 9L, 10L, 11L, 12L, 13L), .Label = c("p00e00d00", "p00e00d11", 
"p00e00d12", "p00e00d13", "p00e00d21", "p00e00d22", "p00e00d23", 
"p00e11d00", "p00e12d00", "p00e13d00", "p00e21d00", "p00e22d00", 
"p00e23d00", "p01e00d00", "p11e00d00", "p11e00d11", "p11e00d12", 
"p11e00d13", "p11e00d21", "p11e00d22", "p11e00d23", "p11e11d00", 
"p11e12d00", "p11e13d00", "p11e21d00", "p11e22d00", "p11e23d00", 
"p12e00d00", "p13e00d00", "p14e00d00", "p21e00d00", "p22e00d00"
), class = "factor"), response = structure(c(2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("2", 
"1"), class = "factor"), n = c(345L, 511L, 583L, 613L, 612L, 
222L, 142L, 531L, 546L, 589L, 636L, 478L, 364L, 313L, 502L, 533L, 
587L, 603L, 385L, 298L, 263L, 518L, 546L, 563L, 593L, 435L, 351L, 
310L, 478L, 579L, 629L, 646L, 357L, 307L, 230L), freq = c(0.408284023668639, 
0.604733727810651, 0.689940828402367, 0.725443786982249, 0.724260355029586, 
0.262721893491124, 0.168047337278107, 0.628402366863905, 0.646153846153846, 
0.697041420118343, 0.752662721893491, 0.565680473372781, 0.430769230769231, 
0.370414201183432, 0.594082840236686, 0.630769230769231, 0.694674556213018, 
0.713609467455621, 0.455621301775148, 0.352662721893491, 0.311242603550296, 
0.61301775147929, 0.646153846153846, 0.666272189349112, 0.701775147928994, 
0.514792899408284, 0.415384615384615, 0.366863905325444, 0.565680473372781, 
0.685207100591716, 0.744378698224852, 0.764497041420118, 0.422485207100592, 
0.363313609467456, 0.272189349112426)), class = c("tbl_df", "tbl", 
"data.frame"), row.names = c(NA, -35L), .Names = c("block", "item", 
"response", "n", "freq"))

有五个块，每个块包含7个项目，某些项目在各个块中具有相同的名称.因此，我可以按以下步骤逐个介绍:

There are five blocks, each block contains 7 items, and some items have the same names across blocks. I can therefore facet by block as follows:

df %>% 
  ggplot(aes(x = item, y = freq)) + 
  geom_bar(stat = "identity", position = "dodge", color = "black") +
  facet_grid(.~block, scales = "free") + 
  coord_cartesian(ylim = c(0, 1), expand = F) +  # need to add expanse = F to prevent zooming away 
  scale_y_continuous(labels = scales::percent) +
  theme(axis.text.x = element_text(angle=45, hjust=1, vjust=1))

我也有矢量，这些矢量说明了每个块中项目应出现的顺序.例如:

I also have vectors which states for each block the order that items should appear in. For example:

block_3_order = c("p11e13d00","p11e12d00", "p11e11d00", "p11e00d00", "p11e21d00", "p11e22d00","p11e23d00")
             )
block_4_order =  c("p00e00d13", "p00e00d12", "p00e00d11", "p00e00d00", "p00e00d21","p00e00d22","p00e00d23")
             )

我试图对项目"因子进行重新排序，但是要获得理想的效果，我需要将数据帧拆分为代表块的子集.否则，我将难以理解如何将因素的排序与项目的ggplot处理整合为各个方面的单一因素.

I tried to reorder the "item" factor, but to get the desired effect I would need to split the dataframe into subsets representing blocks. Otherwise I am having trouble grasping how you can integrate the ordering of factors with the ggplot treatment of item as a single factor across facets.

任何帮助将不胜感激.

推荐答案

要在每个构面中获得不同的自定义轴顺序，可以将每个构面"创建为一个单独的图，然后将它们像布局一样放置在一起.单面图.

To get a different custom axis order in each facet, you can create each "facet" as a separate plot and then lay them out together as if they were a single faceted plot.

library(tidyverse)
#devtools::install_github("baptiste/egg")
library(egg)
library(gridExtra)
library(grid)
theme_set(theme_bw())

首先，创建自定义订单.NULL的那些将在最终绘图中按字母顺序排序.

First, create the custom orderings. The ones that are NULL will just be sorted alphabetically in the final plot.

b.order = list(b1 = NULL,
               b2 = NULL,
               b3 = c("p11e13d00","p11e12d00", "p11e11d00", "p11e00d00", "p11e21d00", "p11e22d00","p11e23d00"),
               b4 = c("p00e00d13", "p00e00d12", "p00e00d11", "p00e00d00", "p00e00d21","p00e00d22","p00e00d23"),
               b5 = NULL)

创建一个绘图列表，每个 block 对应一个.为此，我们将 df 除以 block .为了获得自定义顺序，我们使用 factor 来基于列表 b.order 设置自定义顺序.

Create a list of plots, one for each block. We do this by splitting df by block. To get the custom ordering, we use factor to set the custom order based on the list b.order.

plist = map2(split(df, df$block), b.order,  
     ~ .x %>% group_by(block) %>% 
         mutate(item = factor(item, levels=if(is.null(.y)) sort(unique(item)) else .y)) %>%
       ggplot(aes(x = item, y = freq)) + 
        geom_bar(stat = "identity", position = "dodge", color = "black") +
        facet_grid(.~block, scales = "free") + 
        coord_cartesian(ylim = c(0, 1), expand = F) +  # need to add expanse = F to prevent zooming away 
        scale_y_continuous(labels = scales::percent) +
        theme(axis.text.x = element_text(angle=45, hjust=1, vjust=1),
             plot.margin=margin(b=-5)) +
        labs(x=""))

从除最左边的图以外的所有图中删除y轴标签，标题和刻度线:

Remove y-axis labels, title, and ticks from all but the left-most plot:

plist[2:length(plist)] = plist[2:length(plist)] %>% 
  map(~ .x + theme(axis.text.y=element_blank(),
                   axis.title.y=element_blank(),
                   axis.ticks.y=element_blank()))   

排列图.我们使用 egg 包中的 ggarrange ，以确保绘图面板均具有相同的水平宽度.我们还需要在绘图下添加 Item 标签.但是，即使在 arrangeGrob 内部， ggarrange 也会将绘图打印到输出设备.因此，我们创建对象 p ，清除设备，然后重绘最终图形.

Arrange the plots. We use ggarrange from the egg package in order to ensure that the plot panels all have the same horizontal width. We also need to add the Item label beneath the plot. However, ggarrange prints the plot to the output device, even inside arrangeGrob. So we create the object p, clear the device and then redraw the final plot.

p = arrangeGrob(ggarrange(plots=plist, ncol=length(plist)),
                textGrob("Item"), heights=c(20,1))
grid.newpage()
grid.draw(p)

这篇关于在ggplot中的每个构面内手动订购x轴标签的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！