Variables ordered incorrectly with plot_bivariate

[leosfsm]

I haven't seen any other ticket opened for this issue, so I'm not entirely sure if it is related to something I did wrong.

These are my inputs:

iris_profiles<-iris %>% 
  select(Sepal.Length, Sepal.Width, Petal.Length, Petal.Width) %>% 
  estimate_profiles(n_profiles = 3) %>%
  plot_bivariate()

And my output is the following:

As you can see the variables repeat itself and are incorrectly ordered.

I have tracked down the error to line 37 of R/plot-bivariate.R:

        cors <- apply(expand.grid(vars, vars), 1, paste0, collapse = ".WITH.")[which(upper.tri(diag(vars)))]

The value of cors is the following:

[1] "Sepal.Length.WITH.Sepal.Width" 
[2] "Sepal.Length.WITH.Petal.Length"
[3] "Sepal.Width.WITH.Petal.Length" 
[4] "Sepal.Length.WITH.Petal.Width" 
[5] "Sepal.Width.WITH.Petal.Width"  
[6] "Petal.Length.WITH.Petal.Width" 

If you watch closely the rows are not extracted in the correct order, in this case "Sepal.Length.WITH.Petal.Width" is in position 4 when it should be in position 3. That is a problem because when plot_bivariate() creates the cor_plotlist in [line 183 of R/plot-bivariate.R](https://github.com/data-edu/tidyLPA/blob/master/R/plot-bivariate.R#L183:

    cor_plotlist <- lapply(unique(df_plot$Parameter), function(this_cor){...}

It makes cor_plotlist follow the same order of variable cors , incorrectly ordering the variables and individual plots, resulting in a grid plot like the one shown above.

I made an adjustment to the line 37 of R/plot-bivariate.R:

        cors <- matrix(apply(expand.grid(vars, vars), 1, paste0, collapse = ".WITH."), nrow=length(vars), byrow=TRUE)[which(lower.tri(diag(vars)))]

I don't know if it is the best approach. But my understanding of what R is doing is that it is selecting the matrix elements by columns and not by rows. This way I just transpose the matrix.

[hyunsooseol]

Hi

It looks like the issue that raised above has not been resolved yet. Could you check it for me?

Best Regards
Seol