Plot Tornado-Style Correlation Results Across Variates

Creates a tornado-style plot from mcmodule_corr() results. For each input node, the plot shows all variate-level correlations as small vertical ticks, a black horizontal range line (min to max), a median marker, and a larger marker at the maximum absolute correlation.

Usage

mcmodule_tornado(
  mcmodule = NULL,
  corr_results = NULL,
  output = NULL,
  by_exp = FALSE,
  match_variates = TRUE,
  variates_as_nsv = FALSE,
  print_summary = TRUE,
  progress = FALSE,
  method = c("spearman", "kendall", "pearson"),
  use = "all.obs",
  lim = c(0.025, 0.975),
  colour = "strength"
)

Arguments

mcmodule: (mcmodule object, optional). Module used to compute correlations when corr_results is NULL.
corr_results: (data frame, optional). Output table from mcmodule_corr(). If provided, no new correlation analysis is run. Default: NULL.
output: (character, optional). Output node name. Passed to mcmodule_corr() when corr_results is NULL.
by_exp: (logical). Passed to mcmodule_corr(). Default: FALSE.
match_variates: (logical). Passed to mcmodule_corr(). Default: TRUE.
variates_as_nsv: (logical). Passed to mcmodule_corr(). Default: FALSE.
print_summary: (logical). Passed to mcmodule_corr(). Default: TRUE.
progress: (logical). Passed to mcmodule_corr(). Default: FALSE.
method: (character). Passed to mcmodule_corr(). Default: c("spearman", "kendall", "pearson").
use: (character). Passed to mcmodule_corr(). Default: "all.obs".
lim: (numeric vector). Passed to mcmodule_corr(). Default: c(0.025, 0.975).
colour: (character or logical). Colouring for max absolute points. Default: "strength". If TRUE or "strength", points are coloured by qualitative correlation strength.

Value

A ggplot2 object.

Details

mcmodule_tornado() returns a ggplot object. Use mcmodule_corr() when you need the correlation table; use mcmodule_tornado() when you need a plot object that can be further customised. Interpretation: In the tornado plot each point (one per variate) shows the correlation between that input and the chosen output across the model variates. The coloured point highlights the variate with the maximum absolute correlation for each input and is used to rank inputs. The black point is the median correlation across variates and the black horizontal line shows the range (minimum to maximum) of correlations for that input. The grey horizontal line connects the maximum-absolute point to the zero-correlation vertical line to facilitate interpretation. Use mcmodule_corr() to inspect the numeric per-variate correlations, the plot is designed to give a compact visual summary.