R仿图：base plot画平行坐标图，仿《ggplot2:数据分析与图形艺术》

Posted 2020-12-22 enjoy-respect-9527

　　先放成熟的平行坐标图绘制网址：https://www.r-graph-gallery.com/parallel-plot.html

• 在只用ggplot2包的条件下，《ggplot2:数据分析与图形艺术》的第273页给出了如下画法：

library(ggplot2)
library(tidyr)#数据变形包
pcp_data<-function(df){
      is_numeric<-vapply(df,is.numeric,logical(1))#筛选出数值列
rescale01<-function(x){
      rng<-range(x,na.rm=TRUE)
      (x-rng[1])/(rng[2]-rng[1])
}#将数值列归一化到[0,1]区间
df[is_numeric]<-lapply(df[is_numeric],rescale01)
df$.row<-rownames(df)
tidyr::gather_(df,"variable","value",names(df)[is_numeric])#数据变形
}
pcp<-function(df,...){
      df<-pcp_data(df)
      ggplot(df,aes(variable,value,group=.row))+geom_line(...)
}
pcp(mpg,aes(colour=drv))

 图形如下：

• 用Base plot 仿画，数据集是ggplot2包自带的mpg数据

  library(ggplot2)
  rescale01<-function(x){
        rng<-range(x,na.rm=TRUE)
        (x-rng[1])/(rng[2]-rng[1])
  }
  df<-mpg
  df$col<-ifelse(df$drv=="4","red",ifelse(df$drv=="f","green","purple"))
  is_numeric<-vapply(df,is.numeric,logical(1))
  df[is_numeric]<-lapply(df[is_numeric],rescale01)
  plot.new()
  plot.window(xlim=c(0.8,7),ylim=c(0,1))
  apply(df,1,function(x){lines(1:5,x[c("cty","cyl","displ","hwy","year")],col=x["col"])})
  #平行坐标实际上每一条线就是一个观测，刚好用apply()函数对每行添加线条
  legend(5.5,0.5,legend=c("4","f","r"),col=c("red","green","purple"),cex=1.5,lty=1,title="drv")
  axis(1,at=1:5,labels=c("cty","cyl","displ","hwy","year"))
  axis(2)
  mtext(text="value",side=2,line=2,adj=0.5,cex=1.5)
  mtext(text="variable",side=1,line=2,adj=0.5,cex=1.5)
  abline(v=1:5,lty="dotted")
  

 图形如下：