GGally与pairs相关关系图_史上最全（一）-白红宇

简介对于多个变量之间的相关关系，常常使用相关关系图来可视化，R自带有pairs()函数，可以画相关关系图，但是比较复杂，我们先介绍基于ggplot2的GGally包。等介绍完，再介绍pairs()函数。

1.ggmatrix() ggmatrix()可以将多个ggplot2绘图对象，按照矩阵进行排列。

1.1 矩阵第1列

1library(ggplot2) 2data(tips, package = "reshape") 3 4head(tips) 5 6g1 <- ggplot(tips, aes(x = total_bill, fill = sex)) + 7 geom_density(show.legend = FALSE) 8 9g2 <- ggplot(tips, aes(x = total_bill, fill = sex)) + 10 geom_histogram(position = position_stack(), show.legend = FALSE) + 11 facet_grid(rows = vars(time))# 以time变量行分面 12 13g3 <- ggplot(tips, aes(x = total_bill, y = tip, color = sex)) + 14 geom_point(show.legend = FALSE) 15

1.2 矩阵第2列

1library(ggplot2) 2 3g4 <- ggplot(tips, aes(x = time, y = total_bill, fill = sex)) + 4 geom_boxplot(show.legend = FALSE) 5 6g5 <- ggplot(tips, aes(x = time, fill = sex)) + 7 geom_bar(position = position_stack(), show.legend = FALSE) 8 9g6 <- ggplot(tips, aes(x = tip, fill = sex)) + 10 geom_histogram(position = position_stack(), show.legend = FALSE) + 11 coord_flip() + 12 facet_grid(cols = vars(time)) 13

1.3 矩阵第3列

1library(ggplot2) 2library(dplyr) 3library(tibble) 4 5# 第一个图 6text_1 <- round(cor(tips $total_bill, tips$ tip), 3) 7tips_female <- as.tibble(tips) %>% filter(sex == "Female") %>% as.data.frame() 8tips_male <- as.tibble(tips) %>% filter(sex == "Male") %>% as.data.frame() 9text_2 <- round(cor(tips_female $total_bill, tips_female$ tip), 3) 10text_3 <- round(cor(tips_male $total_bill, tips_male$ tip), 3) 11mytext <- c(text_1, text_2, text_3) 12mytext <- paste0(c("Cor", "Female", "Male"), ":", mytext) 13mytext <- data.frame(text = mytext, 14 x = 5, 15 y = c(6, 4, 2), 16 stringsAsFactors = FALSE) 17 18g7 <- ggplot(data = mytext[-1, ], aes(x = x, y = y, label = text, color = text)) + 19 geom_text(show.legend = F) + 20 geom_text(data = mytext[1,], aes(x = x, y = y, label = text), 21 color = "black") 22 23rm(text_1, tips_female, tips_male, text_2, text_3, mytext) 24 25# 第2个图 26g8 <- ggplot(tips, aes(x = time, y = tip, fill = sex)) + 27 geom_boxplot(show.legend = FALSE) + 28 coord_flip() 29 30# 第3个图 31g9 <- ggplot(tips, aes(x = tip, fill = sex)) + 32 geom_density(show.legend = FALSE) 33

1.4 customLayout合并图形

1library(customLayout) 2# 创建画布 3mylay <- lay_new( 4 mat = matrix(1:9, ncol = 3)) 5 6plot_list <- list(g1, g2, g3, g4, g5, g6, g7, g8, g9) 7 8lay_grid(plot_list, mylay) # ggplot2绘图列表传参，传递到画布mylay 9 10rm(g1, g2, g3, g4, g5, g6, g7, g8, g9, mylay)

1.5 ggmatrix合并图形

1library(GGally) 2 3gg_m <- ggmatrix( 4 plots = plot_list, # 绘图对象列表 5 nrow = 3, ncol = 3, # 行数和列数 6 xAxisLabels = c("Total Bill", "Time of Day", "Tip"), 7 yAxisLabels = c("Total Bill", "Time of Day", "Tip"), 8 byrow = FALSE, # 按列排 9 title = "ggmatrix合并图形" 10) 11 12# 添加主题 13gg_m + theme_bw() 14 15# 提取子集,只能提取其中一个 16gg_m[1,2] 17 18rm(plot_list, gg_m)

2.ggpairs() GGally通过添加几个函数来扩展ggplot2，以降低geom与转换数据组合的复杂性。其中一些功能包括配对图矩阵，散点图矩阵，平行坐标图，生存图，以及绘制网络的几个函数。 2.1 语法及关键参数

语法： 1ggpairs(data, mapping = NULL, columns = 1:ncol(data), title = NULL, 2 upper = list(continuous = "cor", combo = "box_no_facet", discrete = 3 "facetbar", na = "na"), lower = list(continuous = "points", combo = 4 "facethist", discrete = "facetbar", na = "na"), diag = list(continuous = 5 "densityDiag", discrete = "barDiag", na = "naDiag"), params = NULL, ..., 6 xlab = NULL, ylab = NULL, axisLabels = c("show", "internal", "none"), 7 columnLabels = colnames(data[columns]), labeller = "label_value", 8 switch = NULL, showStrips = NULL, legend = NULL, 9 cardinality_threshold = 15, progress = NULL, 10 legends = stop("deprecated"))

关键参数：

mapping, 表示要叠加到x,y上的aes()映射变量，这里是全局映射。 column, 表示选择要绘图的列，可以用变量索引值指定，也可以用变量名指定。 columnLabels, 指定矩阵列名称。 title, xlab, ylab, 表示指定标题和坐标轴名称。 lower,upper,表示指定下三角和上三角的plot类型，列表传参。 diag,表示指定对角线的plot类型，列表传参。 axisLabels, 指定变量名称的显示位置，默认显示在外侧， "internal"则显示在内测，"none"则不显示。 labeller, 表示指定分面标签， switch, 表示指定分面标签位置，与ggplot2:facet_grid中一致，默认在顶部和右侧，若switch = "x",则显示在底部和右侧，若switch = "y"则显示在顶部和左侧，若swith = "both"则显示在底部和左侧。 showStrips, 布尔运算决定是否显示plots的条带，默认NULL只显示顶部和右侧的条带。 TRUE则显示所有的条带，FALSE则不显示所有的条带。 legend, 默认NULL不显示，可以通过theme(legend.position = "bottom")调整图例的位置。有3种指定图例类型的方式：长度为2的数字向量，表示给矩阵所在的行和列增加图例。如c(2,3)表示第2行第3列增加图例。长度为1的数字向量，表示根据矩阵的顺序，给相应的panel添加图例，如legend=3表示给1行第3列增加图例。预先使用grab_legend()提取ggplot2对象的图例，然后指定给legend。 cardinality_threshold, 表示允许因子变量的最大因子水平数量，默认最多15个因子水平。NULL则因子变量不会绘图。 progress, 表示是否显示进度条，默认NULL当超过15个plots时显示进度条, 对绘图结果没有任何影响，不需要关注。 TRUE则显示进度条，FALSE则不显示进度条，也可用ggmatrix_progress()生成进度条，然后指定。

plot类型：通过5个参数控制plot类型：continuous,combo,discret, na, mapping

continuous, 表示如果变量x,y都是连续的，应该是什么plot。对于lower和upper参数：可以是 "point", "smooth","smooth_loess", "density", "cor", "blank"。对于diag参数: 可以是 "densityDiag", "barDiag", "blankDiag" combo, 表示如果变量一个连续，一个离散，应该是什么plot。只能用于lower和upper不能用于diag。离散变量只能计数，不能映射坐标，所以可能存在坐标翻转。可以是 "box", "box_no_facet", "dot", "dot_no_facet", "facethist", "facetdensity", "denstrip", "blank" discrete, 表示2个变量都是离散的，应该是什么plot。对于upper和lower参数：可以是： "facetbar", "ratio", "blank"。对于diag参数: 可以是"barDiag", "blankDiag"。 na, 表示指定变量为na的情况, 对于lower和upper，可以是："na", "blank"。对于diag,可以是 "naDiag"，"blankDiag"。 mapping, 表示aes()映射。若指定mapping参数，则叠加到x,y上去。默认 lower = list(continuous = "point", combo = "facetthist", discrete = "facetbar") 默认 upper = list(continuous = "cor", combo = "box_no_facet", discrete = "box") 默认 diag = list(continuous = "density", discrete = "barDiag")

2.2 column及columnLabels

1library(GGally) 2library(ggplot2) 3 4ggpairs(tips, mapping = aes(color = sex), 5 columns = c("total_bill", "time", "tip"), 6 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Tip(连续变量)"), 7 title = "变量名指定column") 8 9ggpairs(tips, mapping = aes(color = sex), 10 columns = c(1, 6, 2), 11 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Tip(连续变量)"), 12 title = "索引值指定column")

2.3 lower,upper, diag

2.3.1 自定义lower 一个离散变量，lower的discrete参数无效。

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "time", "tip"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Tip(连续变量)"), 4 lower = list( 5 continuous = "cor", 6 combo = "dot_no_facet" # 没有2个离散变量，不需要discrete参数 7 ), 8 upper = list( 9 continuous = "blank", 10 combo = "blank" 11 ), 12 diag = list( 13 continuous = "blankDiag", 14 discrete = "blankDiag" 15 ), 16 title = "自定义lower\n(lower $continuous = \"cor\", lower$ combo = "dot_no_facet")" 17)

两个离散变量，lower的continuous参数无效。

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "time", "sex"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Sex(离散变量)"), 4 lower = list( 5 combo = "dot_no_facet", # 6 discrete = "blank" 7 ), 8 upper = list( 9 combo = "blank", 10 discrete = "blank" 11 ), 12 diag = list( 13 continuous = "blankDiag", 14 discrete = "blankDiag" 15 ), 16 title = "自定义lower\n(lower $combo = \"dot_no_facet\",lower$ discrete = "blank" )" 17)

2.3.2 自定义upper 一个离散变量，upper的discrete参数无效。

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "time", "tip"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Tip(连续变量)"), 4 upper = list( 5 continuous = "density", 6 combo = "dot_no_facet" # 没有2个离散变量，不需要discrete参数 7 ), 8 lower = list( 9 continuous = "blank", 10 combo = "blank" 11 ), 12 diag = list( 13 continuous = "blankDiag", 14 discrete = "blankDiag" 15 ), 16 title = "自定义upper\n(upper $continuous = \"density\", upper$ combo = "dot_no_facet")" 17)

两个离散变量，upper的continuous参数无效。

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "time", "sex"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Sex(离散变量)"), 4 upper = list( 5 combo = "dot_no_facet", # 6 discrete = "ratio" 7 ), 8 lower = list( 9 combo = "blank", 10 discrete = "blank" 11 ), 12 diag = list( 13 continuous = "blankDiag", 14 discrete = "blankDiag" 15 ), 16 title = "自定义upper\n(lower $combo = \"dot_no_facet\",upper$ discrete = "ratio" )" 17)

2.3.3 自定义diag diag没有combo参数。

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "time", "tip"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Tip(连续变量)"), 4 diag = list( 5 continuous = "barDiag", 6 discrete = "blankDiag" # 7 ), 8 lower = list( 9 continuous = "blank", 10 combo = "blank" 11 ), 12 upper = list( 13 continuous = "blank", 14 combo = "blank" 15 ), 16 title = "自定义diag\n(diag $continuous = \"barDiag\", diag$ discrete = "blankDiag")" 17)

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "time", "sex"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Sex(离散变量)"), 4 diag = list( 5 continuous = "barDiag", # 6 discrete = "barDiag" 7 ), 8 lower = list( 9 discrete = "blank", 10 combo = "blank" 11 ), 12 upper = list( 13 discrete = "blank", 14 combo = "blank" 15 ), 16 title = "自定义diag\n(lower $continuous = \"barDiag\",diag$ barDiag = "barDiag" )" 17)

2.3.4 mapping参数 1library(ggplot2) 2library(GGally) 3data(tips, package = "reshape") 4 5ggpairs(tips, 6 columns = c("total_bill", "time", "tip"), 7 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Tip(连续变量)"), 8 title = "无mapping" 9)

1ggpairs(tips, 2 columns = c("total_bill", "time", "tip"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Tip(连续变量)"), 4 lower = list(mapping = aes(color = time)), 5 title = "自定义lower(lower$mapping = "time")" # 局部映射 6)

1ggpairs(tips, 2 columns = c("total_bill", "tip", "size"), 3 columnLabels = c("Total_Bill(连续变量)", "Tip(连续变量)", "Size(连续变量)"), 4 lower = list( 5 continuous = "cor", 6 mapping = aes(color = sex) 7 ), 8 upper = list( 9 continuous = "cor", 10 mapping = aes(color = smoker) 11 ), 12 diag = list( 13 continuous = "barDiag", 14 mapping = aes(color = time) 15 ), 16 title = "自定义lower,upper,diag\n(下三角颜色为sex,上三角颜色为smoker,对角颜色为time)" 17)

2.3.5 同时指定lower,upper,diag 2个连续变量，1个离散变量。

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "tip", "time"), 3 columnLabels = c("Total_Bill(连续变量)", "Tip(连续变量)", "Time(离散变量)"), 4 lower = list( 5 continuous = "cor", 6 combo = "dot_no_facet" # 没有2个离散变量，不需要discrete参数 7 ), 8 upper = list( 9 continuous = "density", 10 combo = "dot_no_facet" # 没有2个离散变量，不需要discrete参数 11 ), 12 diag = list( 13 continuous = "barDiag", 14 discrete = "blankDiag" # 15 ), 16 title = "自定义lower,upper,diag(两个连续变量，一个离散变量)" 17)

1个连续变量，2个离散变量。

1ggpairs(tips, mapping = aes(color = day), 2 columns = c("total_bill", "time", "sex"), 3 columnLabels = c("Total_Bill(连续变量)", "Time(离散变量)", "Sex(离散变量)"), 4 lower = list( 5 combo = "dot_no_facet", # 6 discrete = "blank" 7 ), 8 upper = list( 9 combo = "dot_no_facet", # 10 discrete = "ratio" 11 ), 12 diag = list( 13 continuous = "barDiag", # 14 discrete = "barDiag" 15 ), 16 title = "自定义lower,upper,diag(一个连续变量，两个离散变量)" 17)