R 在facetting上分组选择变量

给定一个数据帧df，如下所示

text <- "
CAR_MODEL,ENGINE_VENDOR,var,value,label
Toyota Corolla,FIAT,Three_Family_Pct,33.98,33.98
Nissan Sunny,PRATNEY,Three_Family_Pct,29.84,29.84
Renault Duster,FIAT,Three_Family_Pct,27.86,27.86
Suzuki Ciaz,FIAT,Three_Family_Pct,26.6,26.6
Renault Duster,FIAT,Single_Family_Pct,0,0
Toyota Corolla,FIAT,Single_Family_Pct,0,0
Nissan Sunny,PRATNEY,Single_Family_Pct,0,0
Suzuki Ciaz,FIAT,Single_Family_Pct,0,0
Suzuki Ciaz,FIAT,Two_Family_Pct,42.37,42.37
Renault Duster,FIAT,Two_Family_Pct,41.53,41.53
Toyota Corolla,FIAT,Two_Family_Pct,36.31,36.31
Nissan Sunny,PRATNEY,Two_Family_Pct,32.27,32.27
Nissan Sunny,PRATNEY,Four_Family_Pct,37.89,37.89
Suzuki Ciaz,FIAT,Four_Family_Pct,31.03,31.03
Renault Duster,FIAT,Four_Family_Pct,30.61,30.61
Toyota Corolla,FIAT,Four_Family_Pct,29.71,29.71
Nissan Sunny,PRATNEY,Mileage,12688.5,12688
Suzuki Ciaz,FIAT,Mileage,11989,11989
Renault Duster,FIAT,Mileage,11132.5,11132
Toyota Corolla,FIAT,Mileage,10357,10357
"
df <- read.table(textConnection(text), sep=",", header = T)

我得到了上面这样的图表。不过，我希望图表中有两处变化

需要

里程

作为最底部的绘图

var

s

Four\u Family\u Pct

、

Three\u Family\u Pct

、

Two\u Family\u Pct

、

Single\u Family\u Pct

，需要以聚集条形图的形式绘制，因为它们属于相同的逻辑属性-它们显示了族类型的百分比

我该怎么做呢

---

@PoGibas-添加了一个用于参考的表示（在excel中制作）-X轴标签需要在那里

像这样的事

library(tidyverse)
p1 <- df %>% 
  mutate(var2=ifelse(var =="Mileage", "Mileage","Family types")) %>% 
  mutate(var=factor(var, levels = levels(df$var)[c(1,3:5,2)])) %>% 
  mutate(label2=ifelse(var =="Mileage", scales::comma(label,digits = 0),paste0(label,"%"))) %>% 
  ggplot(aes(x=CAR_MODEL, y=value, fill=var, label = label2))+ 
     geom_col(aes(col=ENGINE_VENDOR),position = position_dodge(width = 0.9), size=1.2) +
     geom_label(position = position_dodge(width = 0.9) ,show.legend = F)+
  scale_color_manual(values = c(1,2))+
  theme(legend.position="top")
# and the plot using wrap
p1 + facet_wrap( ~ var2, scales = "free_y", ncol = 1) 

对于网格方法，我将删除

aes（col=ENGINE\u VENDOR）

参数

像这样的

library(tidyverse)
p1 <- df %>% 
  mutate(var2=ifelse(var =="Mileage", "Mileage","Family types")) %>% 
  mutate(var=factor(var, levels = levels(df$var)[c(1,3:5,2)])) %>% 
  mutate(label2=ifelse(var =="Mileage", scales::comma(label,digits = 0),paste0(label,"%"))) %>% 
  ggplot(aes(x=CAR_MODEL, y=value, fill=var, label = label2))+ 
     geom_col(aes(col=ENGINE_VENDOR),position = position_dodge(width = 0.9), size=1.2) +
     geom_label(position = position_dodge(width = 0.9) ,show.legend = F)+
  scale_color_manual(values = c(1,2))+
  theme(legend.position="top")
# and the plot using wrap
p1 + facet_wrap( ~ var2, scales = "free_y", ncol = 1) 

---

对于网格方法，我将删除

aes（col=ENGINE\u VENDOR）

参数

不是一个完美的解决方案，但会显示所有需要的信息：

library(ggplot2)

ggplot(df, aes(CAR_MODEL, value)) +
    geom_rect(aes(xmin = CAR_MODEL, xmax = CAR_MODEL,
                  ymin = -Inf, ymax = Inf,
                  color = ENGINE_VENDOR),
                  size = 60) +
    geom_bar(data = subset(df, var != "Mileage"),
             aes(fill = var),
             stat = "identity", position = "dodge", width = 0.6) +
    geom_bar(data = subset(df, var == "Mileage"),
             stat = "identity", width = 0.2, fill = "grey20") +
    facet_wrap(~ var == "Mileage", ncol = 1, scales = "free") +
    labs(x = NULL,
         y = NULL,
         fill = NULL,
         color = NULL) +
    scale_fill_brewer(palette = "Accent") +
    scale_color_brewer(palette = "Paired") +
    theme_classic() +
    theme(strip.background = element_blank(),
          strip.text.x = element_blank(),
          axis.ticks.x = element_blank(),
          legend.position = "top") +
    guides(colour = guide_legend(override.aes = list(fill = brewer.pal(2, "Paired")[1:2], 
                                                     size = 1)))

---

使用

facet\u wrap

查看里程/非里程信息，使用

geom\u rect

查看发动机信息。

不是完美的解决方案，但会显示所有需要的信息：

library(ggplot2)

ggplot(df, aes(CAR_MODEL, value)) +
    geom_rect(aes(xmin = CAR_MODEL, xmax = CAR_MODEL,
                  ymin = -Inf, ymax = Inf,
                  color = ENGINE_VENDOR),
                  size = 60) +
    geom_bar(data = subset(df, var != "Mileage"),
             aes(fill = var),
             stat = "identity", position = "dodge", width = 0.6) +
    geom_bar(data = subset(df, var == "Mileage"),
             stat = "identity", width = 0.2, fill = "grey20") +
    facet_wrap(~ var == "Mileage", ncol = 1, scales = "free") +
    labs(x = NULL,
         y = NULL,
         fill = NULL,
         color = NULL) +
    scale_fill_brewer(palette = "Accent") +
    scale_color_brewer(palette = "Paired") +
    theme_classic() +
    theme(strip.background = element_blank(),
          strip.text.x = element_blank(),
          axis.ticks.x = element_blank(),
          legend.position = "top") +
    guides(colour = guide_legend(override.aes = list(fill = brewer.pal(2, "Paired")[1:2], 
                                                     size = 1)))

---

使用

facet\u wrap

查看里程/非里程信息，使用

geom\u rect

查看发动机信息。

需要以聚集条形图的形式绘制：这会在需要填充堆叠条形图时更改整体图。您同意这样的解决方案吗？@PoGibas-在OPS中添加了一个表示以供参考，但该绘图中没有引擎信息。您希望如何显示它？填充已被使用。@PoGibas-需要引擎信息-但是必须有一种方法来指示条形图中的条形图。@user3206440聚集条形图中颜色的含义是什么？为什么会有8个条形图？需要以聚集条形图的形式绘制：当需要填充的堆叠条形图时，这会更改整体绘图。您同意这样的解决方案吗？@PoGibas-在OPS中添加了一个表示以供参考，但该绘图中没有引擎信息。您希望如何显示它？填充已被使用。@PoGibas-需要引擎信息-但是必须有一种方法来指示条形图中的条形图。@user3206440聚集条形图中颜色的含义是什么？为什么有8条？是的，这很接近-需要在顶部绘图中添加图例键，以便指示哪个条是什么。此外，发动机信息可能会添加到底部绘图中。是的，这已关闭-需要在顶部绘图中添加图例键，以便指示哪个栏是什么。发动机信息也可以添加到底部绘图中。是否可以将值标签添加到每个条形图？是否可以将值标签添加到每个条形图？