Plot development trajectories with optional summary overlay

Visualise loss ratio or related metric trajectories across development time from a Triangle object.

The function supports two display modes:

Raw mode (summary = FALSE): plots cohort-level trajectories coloured by the period variable stored in the Triangle object.
Summary mode (summary = TRUE): plots all cohort trajectories in grey and overlays three summary statistics:
- Mean
- Median
- Weighted mean

Summary statistics are computed from summary.Triangle().

Usage

# S3 method for class 'Triangle'
plot(
  x,
  value_var = "clr",
  summary = FALSE,
  summary_min_n = 5L,
  amount_divisor = 1e+08,
  scales = c("fixed", "free_y", "free_x", "free"),
  theme = c("view", "save", "shiny"),
  ...
)

Arguments

x: An object of class Triangle.
value_var: A single metric to plot. Must be one of: "lr", "clr", "loss", "rp", "margin", "closs", "crp", "cmargin", "loss_prop", "rp_prop", "closs_prop", or "crp_prop".
summary: Logical. If FALSE (default), shows raw cohort trajectories. If TRUE, shows grey cohort trajectories with overlaid summary lines (mean, median, weighted mean). Summary overlay is supported only for "lr" and "clr", and only when the x-axis variable is a development-period variable (for example, elap_m, elap_q, elap_h, elap_y).
summary_min_n: Optional minimum number of observations required for the summary overlay to be considered reliable. When provided and summary = TRUE, a vertical reference line is drawn at the midpoint just before the first development period where n_obs < summary_min_n within each facet. Default is 5.
amount_divisor: Numeric scaling factor used only for y-axis labels of amount variables. Default is 1e8.
scales: Should scales be fixed ("fixed"), free ("free"), or free in one dimension ("free_x", "free_y")?
theme: A string passed to .switch_theme() ("view", "save", "shiny").
...: Additional arguments passed to .switch_theme().

Value

A ggplot object.

Details

The x-axis uses the development variable stored in attr(x, "dev_var"). Cohort lines are grouped by the period variable stored in attr(x, "cohort_var"), and facets are created from attr(x, "group_var").

The loss ratio is defined here as: $$lr = loss / rp$$

where rp denotes risk premium rather than written premium.

The weighted mean is defined as:

lr_wt = sum(loss) / sum(rp)
clr_wt = sum(closs) / sum(crp)

Ratio and proportion metrics are plotted on the original scale and displayed as percentages via y-axis labels. Amount metrics are plotted on the original scale and displayed using y-axis labels scaled by amount_divisor.