如何从 fviz_dist 中生成的有序相异图像中获取矩阵？

Question

I'm trying to obtain the matrix (Ordered dissimilarity matrix) from the ggplot that is generated with the function fviz_dist from factoextra package.

From my original data I generated a distance correlation matrix using dis.cor (I need to use spearman correlation coefficient), as follows:

dist.cor<-get_dist(b, method = "spearman")

c<-round(as.matrix(dist.cor)[1:nrow(b),1:nrow(b)],1)

But this matrix is not ordered so you can't visualize the clusters.

Then, with function fviz_dist I can generate the ggplot of the Distance Correlation Matrix (ordered):

fviz_dist(dist.cor)

But I don't need the image, I need that ordered matrix in a dataframe or matrix format so I can export it using write.csv and open it in excel to work with it.

Thanks!

Answer 1

You can extract the data from the ggplot object:

library(factoextra)
library(tidyr)
dist.cor = get_dist(mtcars,method="spearman")
g = fviz_dist(dist.cor)

The variable names are already factored and ordered:

head(g$data)
             Var1        Var2       value
1     Volvo 142E- Volvo 142E- 0.000000000
2      Fiat X1-9- Volvo 142E- 0.006904570
3       Fiat 128- Volvo 142E- 0.006904570
4 Toyota Corolla- Volvo 142E- 0.006904570
5    Honda Civic- Volvo 142E- 0.009174312
6  Porsche 914-2- Volvo 142E- 0.018254605

head(levels(g$data$Var1))
[1] "Volvo 142E-"     "Fiat X1-9-"      "Fiat 128-"       "Toyota Corolla-"
[5] "Honda Civic-"    "Porsche 914-2-" 

head(levels(g$data$Var2))
[1] "Volvo 142E-"     "Fiat X1-9-"      "Fiat 128-"       "Toyota Corolla-"
[5] "Honda Civic-"    "Porsche 914-2-"

So it's a matter of pivoting it:

head(mat)
mat = pivot_wider(g$data,values_from="value",names_from="Var2")

# A tibble: 6 x 33
  Var1    `Volvo 142E-` `Fiat X1-9-` `Fiat 128-` `Toyota Corolla… `Honda Civic-`
  <fct>           <dbl>        <dbl>       <dbl>            <dbl>          <dbl>
1 Volvo …       0            0.00690     0.00690          0.00690        0.00917
2 Fiat X…       0.00690      0           0                0              0.0254 
3 Fiat 1…       0.00690      0           0                0              0.0254 
4 Toyota…       0.00690      0           0                0              0.0254 
5 Honda …       0.00917      0.0254      0.0254           0.0254         0      
6 Porsch…       0.0183       0.0252      0.0252           0.0252         0.0274