Compute Surprise as ggplot2 Stat — stat_surprise • bayesiansurpriser

This stat computes Bayesian surprise for sf geometries, allowing you to visualize surprise directly in ggplot2.

Usage

stat_surprise(
  mapping = NULL,
  data = NULL,
  geom = "sf",
  position = "identity",
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE,
  models = c("uniform", "baserate", "funnel"),
  signed = TRUE,
  ...
)

Arguments

mapping: Aesthetic mapping created by ggplot2::aes(). Required aesthetics are geometry (from sf) and observed. Optional aesthetics include expected and sample_size.
data: Data (typically an sf object)
geom: Geometry to use (default: "sf")
position: Position adjustment
na.rm: Remove NA values?
show.legend: Show legend?
inherit.aes: Inherit aesthetics from ggplot?
models: Character vector of model types to use. Options: "uniform", "baserate", "gaussian", "sampled", "funnel"
signed: Logical; compute signed surprise?
...: Additional arguments passed to the layer

Value

A ggplot2 layer

Computed Variables

surprise: Bayesian surprise (KL-divergence)
signed_surprise: Signed surprise (if signed = TRUE)

Examples

library(ggplot2)
library(sf)

nc <- st_read(system.file("shape/nc.shp", package = "sf"), quiet = TRUE)

# Basic surprise map - geometry must be mapped explicitly
ggplot(nc) +
  stat_surprise(aes(geometry = geometry, observed = SID74, expected = BIR74)) +
  scale_fill_surprise()