Chord diagram

1 Tương quan biến định tín hnhị phân

library(tidyverse)
library(chorddiag)

set.seed(2009)

dat=data_frame(A=sample(c("Y","N"),500, replace=T,prob=c(0.3,0.7))%>%as.factor(),
               B=sample(c("Y","N"),500, replace=T,prob=c(0.55,0.45))%>%as.factor(),
               C=sample(c("Y","N"),500, replace=T,prob=c(0.2,0.8))%>%as.factor(),
               D=sample(c("Y","N"),500, replace=T,prob=c(0.6,0.4))%>%as.factor(),
               E=sample(c("Y","N"),500, replace=T,prob=c(0.7,0.3))%>%as.factor(),
               X=sample(c("Y","N"),500, replace=T,prob=c(0.8,0.2))%>%as.factor(),
               Y=sample(c("Y","N"),500, replace=T,prob=c(0.4,0.6))%>%as.factor(),
               Z=sample(c("Y","N"),500, replace=T,prob=c(0.35,0.65))%>%as.factor())

head(dat)%>%knitr::kable()

A	B	C	D	E	X	Y	Z
N	Y	Y	Y	N	Y	Y	N
N	Y	Y	N	Y	Y	N	N
N	Y	N	Y	Y	Y	N	N
N	N	Y	Y	Y	N	N	N
N	Y	N	Y	N	Y	N	Y
N	N	Y	Y	Y	Y	N	N

2 Chuẩn bị hàm vẽ chord_diagram

draw_chord_diagram = function(dat = dat, categoric = T) {
  if (categoric == T) {
  cat_list = colnames(dat)
  cor_mat = matrix(nrow = dim(dat)[2], ncol = dim(dat)[2])
  
  for (i in c(1:dim(dat)[2])){
    long_dat = dat %>% gather(cat_list[-i], key = "fact", value = "status")
    cond = long_dat[1]=="Y" & long_dat[3] == "Y"
    filt_long_dat = long_dat[cond,]
    xtb_vect = filt_long_dat %>%group_by(cat_list[1],fact)%>%tally()%>%t()
    cor_mat[i,-i] = xtb_vect[3,]%>%as.integer()
    cor_mat[i,i] = sum(dat[i]=="Y")%>%as.integer()
  }
  
  dimnames(cor_mat) <- list(r = colnames(dat),
                    c = colnames(dat))
  
  chorddiag(cor_mat, 
          groupColors = pals::brewer.spectral(dim(dat)[2]), 
          groupnamePadding = 10,
          showTicks = F)
  }
  else{
    cm = cor(dat)
  diag(cm) = NA
  
  chorddiag(abs(cm),showTicks = F,
            groupColors = pals::brewer.spectral(dim(dat2)[2]))
  }
  }

3 Kết quả

draw_chord_diagram(dat,categoric = T)

4 Chord diagram áp dụng được cả cho biến định lượng

dat2 = data_frame(A= rnorm(500),
                  B= rnorm(500),
                  C= rnorm(500),
                  D= rnorm(500),
                  E= rnorm(500),
                  X= rnorm(500),
                  Y= rnorm(500),
                  Z= rnorm(500))

head(dat2)%>%knitr::kable()

A	B	C	D	E	X	Y	Z
0.5967539	1.0212297	0.1362541	-0.0969490	-0.8733515	-1.0551382	-0.8908629	-1.6135771
0.4787556	-1.0990761	1.4098798	1.3027122	1.0541796	0.3604073	-1.2343853	1.2976629
0.6860399	1.1180576	0.5233555	-0.0043952	1.1081349	0.0964058	-0.0432363	-0.0059015
-0.0895731	0.0577317	-1.2212518	0.2915248	0.3563021	-0.1785445	0.2564587	-0.4844600
-0.3527248	0.6881995	-1.3911750	0.0317707	0.9862375	-2.2138065	0.2614178	1.0838553
-0.2084523	1.8137606	0.7932908	-0.8125831	0.0065936	-0.2444381	-0.2144368	-0.9408870

draw_chord_diagram(dat2,categoric = F)