R 如何熔化包含用于绘图的预先计算的置信区间的数据帧？

我有包含+/-置信区间的数据。我自己从只有+/-CI的二级开放数据集中组装了数据集，因此我无能为力。我知道，使用

ggplot2

绘制多个系列的最快方法是

重塑2

数据帧，我可以很容易地这样做

melt(df, id.vars = c("Year"))

除了将CI列转换为适当的系列之外。现在，我最终想制作一个这样的情节

我用它制作的

ggplot(df, aes(x = Year)) +
  geom_line(aes(y = Total.inflow), color="red") +
  geom_ribbon(aes(ymin = Total.inflow-Total.inflow.CI, ymax = Total.inflow+Total.inflow.CI), colour="red", fill="red", alpha=0.1) +
  geom_line(aes(y = EU.inflow), color="blue") +
  geom_ribbon(aes(ymin = EU.inflow-EU.inflow.CI, ymax = EU.inflow+EU.inflow.CI), colour="blue", fill="blue", alpha=0.1) +
  geom_line(aes(y = ROW.inflow), color="green") +
  geom_ribbon(aes(ymin = ROW.inflow-ROW.inflow.CI, ymax = ROW.inflow+ROW.inflow.CI), colour="green", fill="green", alpha=0.1)

想法

---

工作解决方案

感谢@lukeA为我指出了正确的方法。出于某种原因，他的解决方案产生了一个空的数据帧，但我设法弄明白他试图做什么，并自己找到了一个合理的解决方案

首先，让我们将GDP列与流量数据集分开。我从一开始就怀疑这是必要的，但我相信我可以在策划时过滤掉它。事实证明，把两者分开比较容易。另外，我稍后会将其值标准化，因为数百亿

df <- read.csv('stats.csv', header=T)
gdp <- data.frame(Year = df$Year, GDP = df$GDP/10000000000)
df <- within(df, rm(GDP))

最后，@lukeA的ggplot代码确实生成了他显示的图表

ggplot(final.df, aes(
    x = Year, 
    y = value, 
    ymin = value - conf,
    ymax = value + conf, 
    color = var, 
    fill = var
    )) + 
geom_ribbon(alpha = .2) + 
geom_line() 

---

数据

df <- structure(list(Year = 1991:2014, Total.inflow = c(329L, 268L, 
266L, 315L, 312L, 318L, 327L, 391L, 454L, 479L, 481L, 516L, 511L, 
589L, 567L, 596L, 574L, 590L, 567L, 591L, 566L, 498L, 526L, 632L
), Total.inflow.CI = c(23L, 20L, 19L, 23L, 22L, 25L, 27L, 27L, 
31L, 31L, 30L, 32L, 33L, 40L, 37L, 39L, 40L, 39L, 30L, 31L, 28L, 
27L, 29L, 36L), Total.outflow = c(-285L, -281L, -266L, -238L, 
-236L, -264L, -279L, -251L, -291L, -321L, -309L, -363L, -363L, 
-344L, -361L, -398L, -341L, -427L, -368L, -339L, -351L, -321L, 
-317L, -319L), Total.outflow.CI = c(23L, 21L, 20L, 20L, 19L, 
28L, 24L, 22L, 24L, 27L, 25L, 29L, 32L, 28L, 31L, 34L, 27L, 41L, 
22L, 20L, 22L, 20L, 19L, 22L), UK.inflow = c(93L, 81L, 75L, 91L, 
67L, 75L, 79L, 90L, 92L, 83L, 89L, 74L, 85L, 73L, 82L, 66L, 60L, 
71L, 82L, 84L, 69L, 73L, 70L, 68L), UK.inflow.CI = c(15L, 15L, 
12L, 16L, 13L, 15L, 14L, 15L, 16L, 16L, 16L, 14L, 16L, 12L, 16L, 
14L, 12L, 14L, 13L, 14L, 11L, 11L, 12L, 11L), UK.outflow = c(-142L, 
-146L, -141L, -112L, -130L, -141L, -140L, -121L, -133L, -151L, 
-150L, -172L, -184L, -189L, -175L, -200L, -158L, -159L, -130L, 
-125L, -133L, -131L, -125L, -128L), UK.outflow.CI = c(17L, 16L, 
16L, 14L, 15L, 22L, 19L, 18L, 16L, 18L, 18L, 22L, 22L, 23L, 22L, 
26L, 19L, 22L, 11L, 11L, 12L, 14L, 11L, 13L), EU.inflow = c(60L, 
49L, 48L, 53L, 60L, 74L, 70L, 75L, 64L, 55L, 54L, 57L, 58L, 128L, 
149L, 173L, 189L, 186L, 162L, 171L, 168L, 148L, 193L, 256L), 
    EU.inflow.CI = c(12L, 10L, 8L, 10L, 11L, 14L, 18L, 14L, 16L, 
    13L, 15L, 16L, 17L, 22L, 23L, 26L, 28L, 27L, 19L, 21L, 18L, 
    17L, 20L, 25L), EU.outflow = c(-51L, -39L, -40L, -46L, -38L, 
    -50L, -51L, -52L, -57L, -55L, -50L, -54L, -47L, -45L, -56L, 
    -63L, -66L, -126L, -104L, -92L, -92L, -75L, -78L, -86L), 
    EU.outflow.CI = c(10L, 6L, 7L, 8L, 7L, 13L, 10L, 9L, 13L, 
    12L, 13L, 13L, 16L, 10L, 14L, 15L, 15L, 31L, 16L, 13L, 14L, 
    12L, 12L, 15L), ROW.inflow = c(175L, 138L, 143L, 171L, 185L, 
    169L, 178L, 226L, 298L, 340L, 338L, 385L, 368L, 388L, 336L, 
    358L, 325L, 333L, 323L, 336L, 329L, 277L, 264L, 308L), ROW.inflow.CI = c(13L, 
    10L, 11L, 13L, 15L, 14L, 14L, 17L, 21L, 23L, 20L, 24L, 22L, 
    31L, 25L, 25L, 25L, 25L, 19L, 18L, 19L, 18L, 18L, 24L), ROW.outflow = c(-91L, 
    -96L, -85L, -80L, -69L, -73L, -88L, -78L, -101L, -114L, -109L, 
    -136L, -133L, -109L, -129L, -135L, -117L, -142L, -134L, -122L, 
    -126L, -115L, -114L, -105L), ROW.outflow.CI = c(12L, 12L, 
    10L, 11L, 8L, 10L, 11L, 9L, 14L, 15L, 13L, 15L, 16L, 13L, 
    17L, 16L, 12L, 16L, 9L, 9L, 11L, 9L, 9L, 11L), GDP = c(1142797178130.51, 
    1179659529659.53, 1061388722255.55, 1140489745944.29, 1237561937825.47, 
    1306575663026.52, 1446444007858.55, 1537103345478.64, 1565408509949.85, 
    1554801028899.98, 1535942133294.95, 1680256294964.03, 1943025306122.45, 
    2297889051629.44, 2418941818181.82, 2588077276908.92, 2969733893557.42, 
    2793376838235.29, 2314577036921.64, 2403504326328.8, 2594904662714.31, 
    2630472981169.65, 2712296271989.99, 2990201431078.23)), .Names = c("Year", 
"Total.inflow", "Total.inflow.CI", "Total.outflow", "Total.outflow.CI", 
"UK.inflow", "UK.inflow.CI", "UK.outflow", "UK.outflow.CI", "EU.inflow", 
"EU.inflow.CI", "EU.outflow", "EU.outflow.CI", "ROW.inflow", 
"ROW.inflow.CI", "ROW.outflow", "ROW.outflow.CI", "GDP"), row.names = c(NA, 
-24L), class = "data.frame")

例如，

df

download.file(
  "http://www.sharecsv.com/dl/88f76c7be8ade3a626f474f4857e16f8/stats.csv", 
  tf <- tempfile(), 
  method = "libcurl"
)
library(tidyverse)
df <- read_csv(tf)
inner_join(
  df %>% 
    select(-ends_with("CI")) %>% 
    gather(var, value, -Year),
  df %>% 
    select(Year, ends_with("CI")) %>% 
    setNames(sub("(.*)\\sCI$", "\\1", names(.))) %>% 
    gather(var, conf, -Year),
  by = c("Year", "var")
) %>% 
  ggplot(aes(
    x = Year, 
    y = value, 
    ymin = value - conf,
    ymax = value + conf, 
    color = var, 
    fill = var
  )) + 
  geom_ribbon(alpha = .2) + 
  geom_line() 

download.file(
"http://www.sharecsv.com/dl/88f76c7be8ade3a626f474f4857e16f8/stats.csv", 
tf%
收集（风险值、价值、年份），
df%>%
选择（年份，以“CI”结尾）%>%
集合名称（sub（（.*）\\sCI$，“\\1”，名称（.））%>%
收集（变量、形态、年份），
by=c（“年”、“var”）
) %>% 
ggplot（aes）(
x=年，
y=值，
ymin=value-conf，
ymax=值+形态，
颜色=变量，
fill=var
)) + 
geom_色带（α=0.2）+
geom_线（）

给你


（我正在使用ggplot2的最新开发版本）
谢谢@nrussel。你是如何做出这样的陈述的，这样我下次就可以自己做了；调用
dput（df）
将生成
结构（…）
表达式。谢谢。你能对你的代码稍加评论，解释一下每一部分的作用吗？具体是哪一部分？我宁愿对不清楚的部分进行评论，也不愿写一篇文章即使是对过程的一般描述也可以。我可以计算出每个电话的作用，但我想也许用几行文字来描述这个想法会使你的回答对未来的读者更有用。谢谢，非常感谢。我有没有可能在没有tidyverse的情况下也这样做？似乎它在CRAN中的一些包在编译时遇到了问题……好吧，我找到了它，并设法找到了一个像样的解决方案。我将把它添加到我的问题中，并将你的答案标记为已接受。
download.file(
  "http://www.sharecsv.com/dl/88f76c7be8ade3a626f474f4857e16f8/stats.csv", 
  tf <- tempfile(), 
  method = "libcurl"
)
library(tidyverse)
df <- read_csv(tf)
inner_join(
  df %>% 
    select(-ends_with("CI")) %>% 
    gather(var, value, -Year),
  df %>% 
    select(Year, ends_with("CI")) %>% 
    setNames(sub("(.*)\\sCI$", "\\1", names(.))) %>% 
    gather(var, conf, -Year),
  by = c("Year", "var")
) %>% 
  ggplot(aes(
    x = Year, 
    y = value, 
    ymin = value - conf,
    ymax = value + conf, 
    color = var, 
    fill = var
  )) + 
  geom_ribbon(alpha = .2) + 
  geom_line()