如何使用ggplot2绘制条形图中彼此相邻堆叠的故障类型

Question

我有一个这样的数据框：

ID <- c("A","A","B","B","A","B","B","B","A","A","A","A","B","B","A","A","A","B","B","B")
Measurement <- c("Len","Len","Len","Wid","Ht","Ht","Wid","Len","Ht","Ht"
                 ,"Wid","Ht","Len","Ht","Wid","Len","Wid","Ht","Len","Wid")
STATUS <- c("FAIL","FAIL","FAIL_AVG_HIGH","FAIL","FAIL","FAIL_AVG_HIGH","FAIL#Pts","FAIL","FAIL_AVG_LOW","FAIL"
            ,"FAIL#Pts","FAIL","FAIL_AVG_HIGH","FAIL","FAIL","FAIL_AVG_LOW","FAIL","FAIL_AVG_LOW","FAIL","FAIL#Pts")
df1 <- data.frame(ID,Measurement,STATUS)

我正在创建2列以获取每个类别的失败测量计数并为每个类别创建MEAS_ID列

df2 <- dcast(df1, Measurement+ID~STATUS, value.var="STATUS")
df2$Count <- rowSums(df2[3:6])
df2 <- within(df2, MEAS_ID <- paste(Measurement, ID, sep='_'))

然后，我使用条形图为每个类别绘制帕累托图，如此

windows()
ggplot(df2, aes(reorder(MEAS_ID,Count), Count)) + 
  geom_bar(stat = "identity") + 
  theme(plot.title=element_text(face="bold", size=20)) +
  ggtitle("Count of Failed Measurements") +
  xlab("MEAS_ID") + ylab("Count") +
  coord_flip()

理想情况下，我希望为上面显示的每个类别绘制彼此相邻的故障类型（FAIL，FAIL＃Pts，FAIL_AVG_HIGH，FAIL_AVG_LOW）。我面临的问题是，因为我使用了＆＃34; dcast＆＃34;失败的类型已经从行转移到列中，我不知道如何将它们包含在ggplot中以获得4个子条形图，而不是每个类别的1个条形图。如果不清楚，请告诉我。

请提供您的意见，以帮助我解决这个问题。

Answer 1

您根本不需要使用dcast。相反，你可以在ggplot中使用一些奇特的魔法：

ggplot(df1, aes(x=paste(Measurement, ID), y=..count.., fill=STATUS)) + geom_bar()

如果评论者是正确的，并且听起来像是这样，那么您需要添加一些内容以使每个类别分组，但不会堆叠：

ggplot(df1, aes(x=paste(Measurement, ID), y=..count.., fill=STATUS)) + geom_bar(position = 'dodge')