plot_bivariate function

Create correlation plots for a mixture model

Creates a faceted plot of two-dimensional correlation plots and unidimensional density plots for an object of class 'tidyProfile'.

plot_bivariate(
  x,
  variables = NULL,
  sd = TRUE,
  cors = TRUE,
  rawdata = TRUE,
  bw = FALSE,
  alpha_range = c(0, 0.1),
  return_list = FALSE
)

Arguments

• x: tidyProfile object to plot. A tidyProfile is one element of a tidyLPA analysis.
• variables: Which variables to plot. If NULL, plots all variables that are present in all models.
• sd: Logical. Whether to show the estimated standard deviations as lines emanating from the cluster centroid.
• cors: Logical. Whether to show the estimated correlation (standardized covariance) as ellipses surrounding the cluster centroid.
• rawdata: Logical. Whether to plot raw data, weighted by posterior class probability.
• bw: Logical. Whether to make a black and white plot (for print) or a color plot. Defaults to FALSE, because these density plots are hard to read in black and white.
• alpha_range: Numeric vector (0-1). Sets the transparency of geom_density and geom_point.
• return_list: Logical. Whether to return a list of ggplot objects, or just the final plot. Defaults to FALSE.

Returns

An object of class 'ggplot'.

Examples

# Example 1
iris_sample <- iris[c(1:10, 51:60, 101:110), ] # to make example run more quickly
## Not run:

iris_sample %>%
 subset(select = c("Sepal.Length", "Sepal.Width")) %>%
 estimate_profiles(n_profiles = 2, models = 1) %>%
 plot_bivariate()
## End(Not run)

# Example 2
## Not run:

mtcars %>%
  subset(select = c("wt", "qsec", "drat")) %>%
  poms() %>%
  estimate_profiles(3) %>%
  plot_bivariate()
## End(Not run)

Author(s)

Caspar J. van Lissa