在 R 中使用 colorramp（罗斯林风格气泡图）

Question

我尝试按照以下链接中的答案进行操作，但没有成功。我应该提一下，我正在上入门级数据分析课程，所以我目前的 R 技能是有限的。 How to match vector values with colours from a colour ramp in R?.

无论如何，我的问题是我正在使用 Rosling 风格的气泡图，并且颜色没有正确渐变。该图描绘了有关中国新联盟和 MIDS（军事化的州际争端）的数据。具体来说，圆圈大小表示从 1 到 7 的 MIDS 数。因为区域可能难以辨别，所以我选择使用颜色来更好地区分 MIDS 的#，如图例所示。

图例正确显示了我的配色方案，但是，您可能会注意到较大的圆圈不是深红色，因为它们应该是深红色，也不是所有最小的圆圈都是最浅的颜色。事实上，一些小圆圈和一些大圆圈是深红色的，浅黄色也是如此。我在下面有我的代码和其他详细信息。

我的密码：

library(RColorBrewer)
my.colors99 <-colorRampPalette(c("#ffffcc", "#fd8d3c", "#800026"))(8)
my.design <- my.colors99[ChnData$mids]   

radius1 <- sqrt( ChnData$mids/pi )


symbols ( ChnData$year, ChnData$alliances,
circles= radius1, inches=.35,
xlab = "Year", ylab = "New Alliances", 
main = "China's New Alliances (Circles = MIDS)",
ylim=nrow(ChnData$alliances),
xlim= nrow(ChnData$year), 
bg = my.colors99)

legend("topleft", 
legend=c("0", "1", 
"2", "3", 
"4", "5", 
"6", "7"), title= "# of Mids",
col=my.colors99, pch=19, cex=1, box.col="grey60", 
text.col="grey30")

颜色的一些细节

my.colors99 in R returns:

“#FFFFCC”“#FEDEA2”“#FDBD79”“#FD9D50”

“#EB7838”“#C75032”“#A3282C”“#800026”

ChnData$MIDS in R returns:

“#FFFFCC”“#FEDEA2”“#FFFFCC”“#FFFFCC”“#FFFFCC”“#FFFFCC”“#FFFFCC”
"#FFFFCC" "#FFFFCC" "#FEDEA2" "#FFFFCC" "#FFFFCC" "#FFFFCC" "#FFFFCC"
"#FFFFCC" "#FFFFCC" "#FDBD79" "#FFFFCC" "#FEDEA2" "#FFFFCC" "#FFFFCC"
"#FFFFCC" "#FFFFCC" "#FFFFCC" "#FFFFCC" "#FEDEA2" "#FEDEA2" "#FEDEA2"
"#FDBD79" "#FFFFCC" "#FD9D50" "#FFFFCC" "#FEDEA2" "#FFFFCC" "#FFFFCC"
"#FFFFCC" "#FD9D50" "#FFFFCC" "#FFFFCC" "#FFFFCC" "#FEDEA2" "#FFFFCC"
"#FDBD79" "#FFFFCC" "#FEDEA2" "#FFFFCC" "#FEDEA2" "#FFFFCC" "#FEDEA2"
"#FFFFCC" "#FDBD79" "#FD9D50" "#FEDEA2" "#FFFFCC" "#FD9D50" "#EB7838"
"#FDBD79" "#A3282C" "#FFFFCC" "#EB7838" "#EB7838" "#FDBD79" "#FDBD79"
"#C75032" "#FFFFCC" "#FD9D50" "#C75032" "#FD9D50" "#EB7838" "#FDBD79"
"#FD9D50" "#FFFFCC" "#FFFFCC" "#FEDEA2" "#FD9D50" "#FFFFCC" "#FFFFCC"
"#FFFFCC" "#FDBD79" "#FDBD79" "#FFFFCC" "#FFFFCC" "#FEDEA2" "#FEDEA2"
"#FFFFCC" "#FDBD79" "#FDBD79" "#FDBD79" "#FEDEA2" "#FFFFCC" "#FD9D50"
"#FD9D50" "#EB7838" "#FDBD79" "#FFFFCC" "#FEDEA2" "#FDBD79" "#EB7838"
"#FDBD79" "#EB7838" "#FEDEA2" "#FDBD79" "#FEDEA2" "#FDBD79" "#EB7838"

在我看来，配色方案肯定是关闭的，因为我当然想要最小的圆圈，它代表最低的中间 # 对应于渐变调色板中最亮的颜色，并随着圆圈大小的增加而变暗.希望我提供了足够的细节，但如果需要更多细节，我很乐意回答任何问题。感谢您的帮助！

circles small, color scheme correct with 1-mids but legend isnt

Answer 1

这正是 ggplot2 比基本图形更易于使用的绘图类型

require(ggplot2)
ggplot(ChnData, aes(x = year, y = alliances, col = mids, size = mids)) + 
  geom_point() +
  scale_colour_distiller(palette = "OrRd", direction  = -1) +
  scale_size(range = c(2, 12)) +
  labs(y = "New Alliances", title = "China's New Alliances (Circles = MIDS)")

我怀疑您的代码的问题是my.colors99[ChnData$mids] 没有达到您的预期。当 mids 为零时，它将被丢弃。其余的值将不合适。