使用不同的数字列重塑数据

问题描述

我正在尝试绘制不同的图形，将 HPI 与中值指数进行比较。首先，我在这个例子中有相同的数据，但没有最后三列，我用下面的代码重新调整它，让一列包含中位数，一列包含三个 HPI

 df %>% 

 pivot_wider(names_from = Index,values_from = Value) %>% 
 pivot_longer(starts_with("HPI"),names_to = "Index",values_to = "Value")

现在，我需要其他三列在图形中添加置信区间，但我似乎无法使它起作用。

这是我的数据示例

df <- tibble::tribble(
  ~Index,~Value,~Operation,~Month,~SD,~lower,~upper,"HPI1",0.9,"Sale","01/2020",0.00224245,0.9788445,0.9876349,1.1,"Rent","02/2020",0.00227729,1.0323418,1.0412688,"HPI2",0.89,0.00645641,1.1494260,1.1747352,1.12,"HPI3",0.85,0.00732540,1.1376065,1.1663221,1.22,"Median",0.91,1.02,"03/2020","04/2020",1.1663221)

非常感谢

解决方法

这可能是一种无需使用 pivot_wider 和 pivot_longer 的可行方法：


library(dplyr)
library(lubridate)
library(ggplot2)

# month as date format

df <- 
  df %>% 
  mutate(Month = my(Month))

# All cases of HPI index by Month and Operation to join with Median only data to create a separate data frame for the facetted Median line

df_hpi <- 
  df %>% 
  filter(Index != "Median") %>% 
  select(Index,Month,Operation)

# Median data frame for all facetting cases: Index and operation by month and value
df_med <- 
  df %>% 
  filter(Index == "Median") %>% 
  rename(dataset = Index)%>% 
  right_join(df_hpi,by = c("Month" = "Month","Operation" = "Operation"))

# HPI data only
df1 <- 
  df %>% 
  filter(Index != "Median")
  
# All together in the call to ggplot

ggplot(df1,aes(Month,Value,group = Index))+
  geom_line()+
  geom_line(data = df_med,aes(colour = "Median"))+
  facet_grid(Operation~Index)+
  labs(colour = NULL)+
  theme(legend.position = "bottom")

^{由 reprex package (v2.0.0) 于 2021 年 4 月 13 日创建}

我知道您正在尝试做这样的事情：一条线和一条丝带（置信区间）。但是您的“较低”列的值似乎不正确。

相反，我使用了“SD”列。

library(tidyverse)
library(lubridate)
library(ggplot2)


df <- tibble::tribble(
  ~Index,~Value,~Operation,~Month,~SD,~lower,~upper,"HPI1",0.9,"Sale","01/2020",0.00224245,0.9788445,0.9876349,1.1,"Rent","02/2020",0.00227729,1.0323418,1.0412688,"HPI2",0.89,0.00645641,1.1494260,1.1747352,1.12,"HPI3",0.85,0.00732540,1.1376065,1.1663221,1.22,"Median",0.91,1.02,"03/2020","04/2020",1.1663221)

df %>%
  mutate(Month = my(Month),ymin = Value - 2*SD,ymax = Value + 2*SD) %>%
  ggplot() +
  geom_line(aes(
    x = Month,y = Value,group = Operation,color = Operation
  )) +
  geom_ribbon(aes(
    x = Month,ymin = ymin,ymax = ymax,fill = Operation
  ),alpha = 0.3) +
  facet_wrap(~ Index) +
  theme_light() +
  labs(x = NULL,y = NULL)

enter image description here

r r reshape