Create correlation plot — create_correlation

Renders a correlation heatmap for small matrices, OR a top-N effect-size lollipop chart for large matrices. An earlier corrplot heatmap was unconditional, producing a 14,280x14,280 PNG (4.8 MB, browser-illegible) on a 228-variable run. Above the max_inline_vars threshold the function now switches to a ggplot2 horizontal lollipop showing the top-N pairs ranked by absolute correlation, with significance encoded by point color.

Usage

create_correlation_plot(
  results,
  output_path,
  methodology_mode = NULL,
  run_id = NULL,
  max_inline_vars = 30L,
  excluded_pairs = NULL
)

Arguments

results: CorrelationResults from calculate_correlations()
output_path: File path for PNG output
methodology_mode: Optional character. When supplied, adds a footer caption identifying the methodology mode + run.
run_id: Optional character: run identifier.
max_inline_vars: Integer; correlation matrices with more variables than this render as a top-N lollipop instead of a heatmap. Default 30L.
excluded_pairs: Optional data frame (the extract_significant() result) carrying var1, var2, and excluded_from_findings. Pairs flagged TRUE (analyst-internal / circular, e.g. an affect instrument x a theme-membership column) are shown but never presented as a significant finding – the heatmap blanks them (and discloses the count) and the lollipop marks them "excluded (circular)". NULL (default) reproduces the pre-#2b plot exactly.