将多个阴影/矩形添加到ggplot2图形

Question

我正在尝试在ggplot2图表上添加多个阴影/矩形。在这个可重复的例子中，我只添加了3个，但我可能需要使用完整数据加起来一百个。

以下是我的原始数据的子集 - 在名为temp的数据框中 - dput位于问题的底部：

     Season tier  group value
NA       NA   NA   <NA>    NA
99     1948    2   Wins    20
101    1948    2 Losses    17
NA.1     NA   NA   <NA>    NA
NA.2     NA   NA   <NA>    NA
104    1951    2   Wins    21
106    1951    2 Losses    18
107    1952    2   Wins    23
109    1952    2 Losses    18
110    1953    2   Wins    25
112    1953    2 Losses    18
113    1954    2   Wins    26
115    1954    2 Losses    19
116    1955    2   Wins    26
118    1955    2 Losses    19
119    1956    2   Wins    26
121    1956    2 Losses    20
NA.3     NA   NA   <NA>    NA
123    1958    1   Wins    27
125    1958    1 Losses    20
126    1959    1   Wins    27
128    1959    1 Losses    21
129    1960    1   Wins    28
131    1960    1 Losses    21
132    1961    1   Wins    30
134    1961    1 Losses    21
135    1962    1   Wins    30
137    1962    1 Losses    23
138    1963    1   Wins    31
140    1963    1 Losses    23
141    1964    1   Wins    32
143    1964    1 Losses    23
144    1965    1   Wins    34
146    1965    1 Losses    23
NA.4     NA   NA   <NA>    NA

我可以像这样制作一个ggplot：

p <- ggplot(temp, aes(Season,value, color=group)) + geom_point(size=4, shape=19) +
  scale_color_manual(values=c("red", "gray55"))
p

现在，我想添加阴影。每个阴影将以季节列中开始一系列日期的季节开始，并以季节列中日期行中的最后一个日期/季节结束。最后，每个阴影都应该由“等级”着色。变量。绿色为＆＃39;等级== 2＆＃39;和蓝色为＆＃39; tier == 1＆＃39;。

我已经使用了rle函数加上一些额外的脚本来提取Season变量中一系列NAs结束后的第一个观察结果，以及在运行中的第一个NA之前获取观察结果。来港。然后我分别添加和减去0.5，这给了我另一个数据帧（tempindex），如下所示：

#    xmin   xmax ymin ymax
#5 1947.5 1948.5 -Inf  Inf
#6 1950.5 1956.5 -Inf  Inf
#7 1957.5 1965.5 -Inf  Inf

我可以像这样手动添加阴影：

t2.rect1 <- data.frame (xmin=1947.5, xmax=1948.5, ymin=-Inf, ymax=Inf)
t2.rect2 <- data.frame (xmin=1950.5, xmax=1956.5, ymin=-Inf, ymax=Inf)
t1.rect1 <- data.frame (xmin=1957.5, xmax=1965.5, ymin=-Inf, ymax=Inf)

p + 
  geom_rect(data=t2.rect1, aes(xmin=xmin, xmax=xmax, ymin=ymin, ymax=ymax), fill="green", alpha=0.1, inherit.aes = FALSE) +
  geom_rect(data=t2.rect2, aes(xmin=xmin, xmax=xmax, ymin=ymin, ymax=ymax), fill="green", alpha=0.1, inherit.aes = FALSE) +
  geom_rect(data=t1.rect1, aes(xmin=xmin, xmax=xmax, ymin=ymin, ymax=ymax), fill="blue", alpha=0.1, inherit.aes = FALSE)

给出了所需的输出：

然而，当我有很多阴影/矩形应用时，我显然不想手动写出这一百次。我想看看我能否以自动化的方式做到这一点。我尝试了一个for循环（不试图逐层调整颜色）非常不成功......

grect <-vector("list", nrow(indextemp)) #vector for storing geom_rects

 for (i in 1:nrow(indextemp)){
 grect[[i]] <- geom_rect(data=temp[i], aes(xmin=xmin, xmax=xmax, ymin=ymin, ymax=ymax), alpha=0.1, inherit.aes = FALSE)
 }

这显然不是首发。我想知道是否有人有任何想法？

对于临时数据框

dput：

structure(list(Season = c(NA, 1948L, 1948L, NA, NA, 1951L, 1951L, 
1952L, 1952L, 1953L, 1953L, 1954L, 1954L, 1955L, 1955L, 1956L, 
1956L, NA, 1958L, 1958L, 1959L, 1959L, 1960L, 1960L, 1961L, 1961L, 
1962L, 1962L, 1963L, 1963L, 1964L, 1964L, 1965L, 1965L, NA), 
    tier = c(NA, 2L, 2L, NA, NA, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
    2L, 2L, 2L, 2L, 2L, NA, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, NA), group = structure(c(NA, 
    1L, 3L, NA, NA, 1L, 3L, 1L, 3L, 1L, 3L, 1L, 3L, 1L, 3L, 1L, 
    3L, NA, 1L, 3L, 1L, 3L, 1L, 3L, 1L, 3L, 1L, 3L, 1L, 3L, 1L, 
    3L, 1L, 3L, NA), .Label = c("Wins", "Draws", "Losses"), class = "factor"), 
    value = c(NA, 20L, 17L, NA, NA, 21L, 18L, 23L, 18L, 25L, 
    18L, 26L, 19L, 26L, 19L, 26L, 20L, NA, 27L, 20L, 27L, 21L, 
    28L, 21L, 30L, 21L, 30L, 23L, 31L, 23L, 32L, 23L, 34L, 23L, 
    NA)), .Names = c("Season", "tier", "group", "value"), row.names = c("NA", 
"99", "101", "NA.1", "NA.2", "104", "106", "107", "109", "110", 
"112", "113", "115", "116", "118", "119", "121", "NA.3", "123", 
"125", "126", "128", "129", "131", "132", "134", "135", "137", 
"138", "140", "141", "143", "144", "146", "NA.4"), class = "data.frame")

对于tempindex数据帧

dput：

structure(list(xmin = c(1947.5, 1950.5, 1957.5), xmax = c(1948.5, 
1956.5, 1965.5), ymin = c(-Inf, -Inf, -Inf), ymax = c(Inf, Inf, 
Inf)), .Names = c("xmin", "xmax", "ymin", "ymax"), row.names = 5:7, class = "data.frame")

Answer 1

最好只使用一个图层，并使用合适的映射

tempindex <- transform(tempindex, 
                       id = 1:3,
                       tier = c(1,1,2))


ggplot(temp, aes(Season,value, color=group)) + 
  geom_rect(data=tempindex, inherit.aes=FALSE,
            aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax,
                group=id, fill = factor(tier)), alpha=0.2)+
  geom_point(size=4, shape=19) +
  scale_color_manual(values=c("red", "gray55"))+
  scale_fill_manual(values=c("green", "blue")) +
  guides(fill="none")

Answer 2

另一种选择（避免创建辅助data.frames更直接）可能是这样：

ggplot(tmp, aes(Season,value, color = group)) + geom_point(size = 4, shape = 19) +
  scale_color_manual(values = c("red", "gray55")) + 
  annotate("rect", xmin = c(1947.5, 1950.5, 1957.5), xmax = c(1948.5, 1956.5, 1965.5), 
           ymin = -Inf, ymax = Inf, alpha = .1, fill = c("green", "green", "blue"))