Skip to contents

Build the main incubation-period analysis figure

Source: R/ploting_utils.R
plot_main_figure.Rd

Produces a multi-panel figure with one tile per pathogen. Each tile shows the posterior predictive CDF (95 % credible ribbon + median line) for the best-fitting distribution selected by LOO-based model weights (pseudo-Bayes factors; see compute_pathogen_model_bayes_factors()). Dashed segments at the P50 and P95 marks span the 95 % posterior uncertainty of those quantiles.

When subgroup analyses are present (e.g. country or variant strata), their median CDFs are overlaid as thin coloured lines without ribbons; the overall "filtered" analysis remains the primary visual element. The chosen distribution is labelled in the bottom-right corner together with the posterior median of the mean delay and shape parameter(s). A model-weight annotation (\(w\)) is shown in the top-left when one model clearly dominates (\(w \ge 0.75\)).

Usage

plot_main_figure(
  all_results,
  pathogens = NULL,
  pathogen_labels = NULL,
  x_max = NULL,
  n_draws = 500L,
  model_weights = NULL,
  show_subgroups = TRUE,
  max_subgroups = 8L,
  ncol = 4L,
  base_size = 9,
  show_title = TRUE,
  pathogen_highlights = NULL
)

Arguments

all_results

Nested list produced by analysis/main_analysis.R.

pathogen_labels

Optional named character vector overriding display labels (e.g. c(COVID_19 = "COVID-19")).

x_max

Upper limit of the x-axis (days). NULL (default) derives each panel's limit automatically as the nearest 5-day step at or above the 99th percentile of the posterior predictive CDF (minimum 10 days). Alternatively, supply a single numeric applied to all panels, or a named numeric vector with per-pathogen overrides plus an optional "default" entry.

n_draws

Integer. Posterior draws for compute_predictive_cdf() for the overall analysis. Subgroup CDFs use ceiling(n_draws / 2).

model_weights

Pre-computed output from compute_pathogen_model_bayes_factors(). If NULL, weights are computed internally — this is slow (1–2 hours for a full set of pathogens). Pre-compute and cache this object.

show_subgroups

Logical. Overlay subgroup median CDFs (default TRUE).

max_subgroups

Integer. Maximum subgroup lines per panel (default 8).

ncol

Integer. Columns in the assembled figure (default 4).

base_size

Numeric. Base font size for ggplot2::theme_bw() (default 9).

Value

A patchwork ggplot object.

See also

compute_pathogen_model_bayes_factors()