如何对漏斗图中的条形重新排序

Posted 2023-02-16

【中文标题】如何对漏斗图中的条形重新排序

【英文标题】：how to reorder bars in a funnel chart

【发布时间】：2021-09-01 11:11:52

【问题描述】：

我想画一个漏斗图，但是所有的条形图都没有正确排序。

funnel_dt <- read.csv2("https://raw.githubusercontent.com/Deborah-Jia/Complete_Analysis_da2/main/funnel_dt.csv")

funnel_dt %>% ggplot(aes(x = district, y = N, fill = covid)) +   # Fill column
  geom_bar(stat = "identity", width = .6) +   # draw the bars
  scale_y_continuous(breaks = brks, labels = lbls) + # Labels 
  scale_x_continuous(breaks= seq(1,23,1) ,labels=paste0("district ", as.character(seq(1, 23, 1)))) +
  coord_flip() +  # Flip axes
  labs(title="") +
  theme_tufte() +  # Tufte theme from ggfortify
  theme(plot.title = element_text(hjust = .5), 
        axis.ticks = element_blank()) +   # Centre plot title
  scale_fill_brewer(palette = "Dark2")  # Color palette

剧情比较混乱。

如何将最长的条（11 区、13 区等）放在底部？我试过reorder，但它不起作用

【问题讨论】：

你是如何尝试reorder的？由于您有正数和负数但想按总长度排序，我认为reorder(district, N, function(x) sum(abs(x))) 会起作用。

我在coord_flip() + 之后使用reorder(N, district)，它说Discrete value supplied to continuous scale。我也试过你的ggplot(aes(x = reorder(district, N, function(x) sum(abs(x))), y = N, fill = covid))同样的错误。我说错了吗？

那个问题是你有 scale_x_continuous 但你的 x 值（区）是离散的，而不是连续的。看我的回答。

【参考方案1】：

您有正数和负数，但想按总长度排序，所以我们需要对绝对值求和并按此排序：

funnel_dt <- read.csv2("https://raw.githubusercontent.com/Deborah-Jia/Complete_Analysis_da2/main/funnel_dt.csv")

funnel_dt %>% 
  mutate(district = reorder(district, N, function(x) -sum(abs(x)))) %>%
  ggplot(aes(x = district, y = N, fill = covid)) +   # Fill column
  geom_bar(stat = "identity", width = .6) +   # draw the bars
  #scale_y_continuous(breaks = brks, labels = lbls) + # Labels 
  scale_x_discrete(labels = function(d) paste("District", d)) +
  coord_flip() +  # Flip axes
  labs(title="") +
  #theme_tufte() +  # Tufte theme from ggfortify
  theme(plot.title = element_text(hjust = .5), 
        axis.ticks = element_blank()) +   # Centre plot title
  scale_fill_brewer(palette = "Dark2")  # Color palette

我注释掉了 y 刻度，因为我没有 brks 和 labels，但是您的 x 刻度（翻转为垂直）是离散的，而不是连续的 - 并且要小心覆盖标签和手动设置中断，可以覆盖排序。我留下了中断并使用labels 的匿名函数粘贴到“District”上。

【参考方案2】：

我使用pivot_wider 将covid 列分成两列

funnel_dt1 <- tidyr::pivot_wider(funnel_dt, names_from = 'covid', values_from = 'N')
funnel_dt1 <- funnel_df1[order(funnel_df1$after),]

funnel_dt1$district <- paste(rep('district ',nrow(funnel_dt1)), funnel_dt1$district)

funnel_df1 %>% ggplot(aes(x = reorder(district, -after))) +   # Fill column
  geom_bar(aes(y = before), fill = 'green', stat = "identity", width = .6) +
  geom_bar(aes(y = after), fill = 'red', stat = "identity", width = .6) +
  scale_x_discrete(labels = function(x) paste("District", x)) +
  coord_flip() +  # Flip axes
  labs(title="") +
  # theme_tufte() +  # Tufte theme from ggfortify
  theme(plot.title = element_text(hjust = .5), 
        axis.ticks = element_blank())    # Centre plot title