Question

请查看我的可重复示例和结果图表。

X = t(USArrests)

plot_color_clust = function(X,N=N,
cols=c("red","blue", "orange", "darkgreen","green","yellow","grey","black","white") 
                            ){
  library(ggplot2)
  library(gridExtra)
  library(gtable)
  library(scales)
  library(ggdendro)
  library(grid)
  library(plyr)

  if(N>length(cols)) stop("N too big. Not enough colors in cols.")
  if(N>ncol(X)) stop("N too big. Not enough columns in data.")

  fit = ClustOfVar::hclustvar(X.quanti = X)
  dd.row = as.dendrogram(fit)
  ddata_x <- dendro_data(dd.row)
  temp = cutree(fit,k=N)
  lab <- ggdendro::label(ddata_x)

  x=c()
  for(i in 1:nrow(lab)){
    x[i]=    paste( "clust", as.vector(temp[ lab$label[i]==names(temp) ])   ,sep="")
  }

  lab$group <- x

  p1 <- ggplot(segment(ddata_x)) + 
    geom_segment(aes(x=x, y=y, xend=xend, yend=yend))+coord_flip()+
    geom_text(data=lab,
              aes(label=label, x=x, y=0, colour=group),hjust=1) +
    theme(legend.position="none",
          axis.title.y=element_blank(),
          axis.title.x=element_blank(),
          axis.text.x = element_text(angle = 0, hjust = 0),
          axis.title.x = element_text(angle = 0, hjust = 0))+
    theme(axis.text = element_blank(), axis.title = element_blank(), 
          axis.ticks = element_blank(), axis.ticks.margin = unit(0, "lines"), 
          axis.ticks.length = unit(0, "cm"))+
    scale_colour_manual(values=cols)+coord_flip()+
    scale_y_continuous(limits = c(-0.1, 2.1))

  df2<-data.frame(cluster=cutree(fit,N),states=factor(fit$labels,levels=fit$labels[fit$order]))
  df3<-ddply(df2,.(cluster),summarise,pos=mean(as.numeric(states)))
  p2 = ggplot(df2,aes(states,y=1,fill=factor(cluster)))+geom_tile()+
    scale_y_continuous(expand=c(0,0))+
    theme(axis.title=element_blank(),
          axis.ticks=element_blank(),
          axis.text=element_blank(),
          legend.position="none")+coord_flip()+
    geom_text(data=df3,aes(x=pos,label=cluster))+
    scale_fill_manual(name = "This is my title", values = cols)

  gp1<-ggplotGrob(p1)
  gp2<-ggplotGrob(p2)  
  maxHeight = grid::unit.pmax(gp1$heights[2:5], gp2$heights[2:5])
  gp1$heights[2:5] <- as.list(maxHeight)
  gp2$heights[2:5] <- as.list(maxHeight)
  #grid.arrange(gp2, gp1, ncol=2,widths=c(1/6,5/6))
  R = arrangeGrob(gp2,gp1,ncol=2,widths=c(1/6,5/6))
  R

}

plot_color_clust(X,6)

问题：

这两个部分（左侧颜色图块和右侧聚类树）具有不一致的高度。我们如何调整他们的高度以便彼此匹配？
如何让右侧的树更短，以便状态名称（聚类主题）可以有更多空间可以完全显示？
有没有办法让这两个部分之间的空白变小？

您对代码的调整表示赞赏。感谢。

Answer 1

一个重大变化：我从gp2中提取绘图面板，然后将其插入gp1的第2列，而不是匹配两个图表的高度。结果gp2周围没有边距，因此，部分照顾你的观点3。

关于第2点：扩展轴的极限以腾出标签的空间。（参见下面代码中的第2点）。点2和3的参数通过反复试验设定。调整一个参数意味着需要调整另一个参数。

关于第1点：使用exapnd的加法分量展开轴，将轴的每一端加半个单位（参见下面代码中的第1点）。

次要编辑：更新为ggplot2 2.2.0和R 3.3.2
不推荐使用axis.ticks.margin

X = t(USArrests)

plot_color_clust = function(X, N = N,
 #  cols=c("red","blue", "orange", "darkgreen","green","yellow","grey","black","white")
   cols = rainbow(N)   # Easier to pick colours
  ){

  library(ggplot2)
  library(gtable)
  library(grid)
  library(ggdendro)
  library(plyr)

  if(N > length(cols)) stop("N too big. Not enough colors in cols.")
  if(N > ncol(X)) stop("N too big. Not enough columns in data.")

  fit = ClustOfVar::hclustvar(X.quanti = X)
  dd.row = as.dendrogram(fit)
  ddata_x <- dendro_data(dd.row)
  temp = cutree(fit, k = N)
  lab <- ggdendro::label(ddata_x)

  x = c()
  for(i in 1:nrow(lab)){
    x[i] = paste("clust", as.vector(temp[lab$label[i] == names(temp)]), sep = "")
  }

  lab$group <- x

  p1 <- ggplot(segment(ddata_x)) + 
    geom_segment(aes(x = x, y = y, xend = xend, yend = yend)) + 
    geom_text(data = lab, aes(label = label, x = x, y = -.05, colour = group),  # y = -.05 adds a little space between label and tree
              size = 4, hjust = 1) +
    scale_x_continuous(expand = c(0, .5)) +   # 1. Add half a unit to each end of the vertical axis
    expand_limits(y = -0.4) +   # 2. Make room for labels 
    theme_classic() + 
    scale_colour_manual(values = cols) + 
    coord_flip() +
    theme(legend.position = "none", axis.line = element_blank(),
          axis.text = element_blank(), axis.title = element_blank(), 
          axis.ticks = element_blank(), 
          axis.ticks.length = unit(0, "cm")) 

  df2 <- data.frame(cluster = cutree(fit, N), 
      states = factor(fit$labels, levels = fit$labels[fit$order]))
  df3 <- ddply(df2, .(cluster),summarise,pos=mean(as.numeric(states)))
  p2 <- ggplot(df2, aes(states, y = 1, 
                    fill = factor(as.character(cluster)))) +   # 'as.character' - so that colours match with 10 or more clusters
    geom_tile() +
    scale_y_continuous(expand = c(0, 0)) + 
    scale_x_discrete(expand = c(0, 0)) +
    coord_flip() +
    geom_text(data = df3,aes(x = pos, label = cluster, size = 12)) +
    scale_fill_manual(values = cols)

  gp1 <- ggplotGrob(p1)  # Get ggplot grobs
  gp2 <- ggplotGrob(p2)  

  gp2 <- gp2[6, 4]      # 3. Grab plot panel only from tiles plot (thus, no margins)
  gp1 <- gtable_add_grob(gp1, gp2, t = 6, l = 2, name = "tiles")  # 3. Insert it into dendrogram plot
  gp1$widths[2] = unit(1, "cm")  # 3. Set width of column containing tiles

  grid.newpage()
  grid.draw(gp1)
}

plot_color_clust(X, 6)

R：ggplot对聚类摘要的轻微调整

1 个答案: