Individual Conditional Expectation (ICE) Plot — ggplot2

An Individual Conditional Expectation (ICE) plot visualizes how the predicted outcome of a machine learning model changes for each individual observation as a single feature varies across its range. Unlike partial dependence plots (PDP) that show the average marginal effect, ICE plots display one line per observation, revealing heterogeneous effects, feature interactions, and subgroup-specific behaviors that would be hidden by averaging. This makes ICE plots essential for detecting when a feature's effect varies across the population.

Individual Conditional Expectation (ICE) Plot rendered with ggplot2

Renders

R source (ggplot2)

#' anyplot.ai
#' ice-basic: Individual Conditional Expectation (ICE) Plot
#' Library: ggplot2 3.5.1 | R 4.4.1
#' Quality: 92/100 | Created: 2026-08-17

library(ggplot2)
library(dplyr)
library(tidyr)
library(scales)
library(ragg)

set.seed(42)

# --- Theme tokens -------------------------------------------------------
THEME       <- Sys.getenv("ANYPLOT_THEME", "light")
PAGE_BG     <- if (THEME == "light") "#FAF8F1" else "#1A1A17"
INK         <- if (THEME == "light") "#1A1A17" else "#F0EFE8"
INK_SOFT    <- if (THEME == "light") "#4A4A44" else "#B8B7B0"
IMPRINT_PALETTE <- c("#009E73", "#C475FD", "#4467A3", "#BD8233",
                     "#AE3030", "#2ABCCD", "#954477", "#99B314")

# --- Data -----------------------------------------------------------------
# ICE curves from a house-price model: predicted sale price as home size
# varies, one curve per house. Two age cohorts reveal a feature interaction —
# older homes plateau above a size threshold, newer homes keep climbing.
n_houses <- 100
n_grid <- 70

age_levels <- c("Newer build (<15 yr)", "Older build (15+ yr)")
house_age <- factor(
  sample(age_levels, n_houses, replace = TRUE, prob = c(0.45, 0.55)),
  levels = age_levels
)

sqft_grid <- seq(800, 3500, length.out = n_grid)

base_price      <- rnorm(n_houses, mean = 180000, sd = 22000)
price_per_sqft  <- ifelse(house_age == "Newer build (<15 yr)",
                           rnorm(n_houses, mean = 148, sd = 14),
                           rnorm(n_houses, mean = 96, sd = 18))
plateau_sqft    <- ifelse(house_age == "Older build (15+ yr)",
                           rnorm(n_houses, mean = 2200, sd = 150), Inf)
wiggle_amplitude <- rnorm(n_houses, mean = 0, sd = 9000)
wiggle_phase     <- runif(n_houses, 0, 2 * pi)

house_params <- tibble::tibble(
  observation_id   = seq_len(n_houses),
  house_age        = house_age,
  base_price       = base_price,
  price_per_sqft   = price_per_sqft,
  plateau_sqft     = plateau_sqft,
  wiggle_amplitude = wiggle_amplitude,
  wiggle_phase     = wiggle_phase
)

ice_df <- expand_grid(observation_id = seq_len(n_houses), feature_value = sqft_grid) %>%
  left_join(house_params, by = "observation_id") %>%
  mutate(
    effective_sqft = pmin(feature_value, plateau_sqft) +
      0.18 * pmax(feature_value - plateau_sqft, 0),
    prediction = base_price + price_per_sqft * effective_sqft +
      wiggle_amplitude * sin(feature_value / 650 + wiggle_phase)
  ) %>%
  select(observation_id, house_age, feature_value, prediction)

pdp_df <- ice_df %>%
  group_by(feature_value) %>%
  summarize(prediction = mean(prediction), .groups = "drop")

observed_sqft <- tibble::tibble(
  feature_value = pmin(pmax(rnorm(n_houses, 1900, 480), 800), 3500)
)

# --- Plot -------------------------------------------------------------------
p <- ggplot() +
  geom_line(
    data = ice_df,
    aes(x = feature_value, y = prediction, group = observation_id, color = house_age),
    alpha = 0.12, linewidth = 0.35
  ) +
  geom_rug(
    data = observed_sqft,
    aes(x = feature_value),
    sides = "b", color = INK_SOFT, alpha = 0.35, linewidth = 0.3
  ) +
  geom_line(
    data = pdp_df,
    aes(x = feature_value, y = prediction),
    color = INK, linewidth = 1.6
  ) +
  scale_color_manual(values = IMPRINT_PALETTE[1:2]) +
  scale_y_continuous(labels = label_dollar(scale = 1e-3, suffix = "K")) +
  guides(color = guide_legend(override.aes = list(alpha = 1, linewidth = 2))) +
  labs(
    title = "ice-basic · r · ggplot2 · anyplot.ai",
    x = "Home Size (sq ft)",
    y = "Predicted Sale Price",
    color = "House Age"
  ) +
  theme_minimal(base_size = 8) +
  theme(
    plot.background     = element_rect(fill = PAGE_BG, color = PAGE_BG),
    panel.background    = element_rect(fill = PAGE_BG, color = NA),
    panel.grid.major.x  = element_blank(),
    panel.grid.minor    = element_blank(),
    panel.grid.major.y  = element_line(color = INK_SOFT, linewidth = 0.2),
    axis.line           = element_line(color = INK_SOFT),
    axis.title          = element_text(color = INK, size = 10),
    axis.text           = element_text(color = INK_SOFT, size = 8),
    plot.title          = element_text(color = INK, size = 12),
    legend.background   = element_blank(),
    legend.key          = element_blank(),
    legend.text         = element_text(color = INK_SOFT, size = 8),
    legend.title        = element_text(color = INK, size = 10)
  )

# --- Save -------------------------------------------------------------------
ggsave(
  filename = sprintf("plot-%s.png", THEME),
  plot     = p,
  device   = ragg::agg_png,
  width    = 8,
  height   = 4.5,
  units    = "in",
  dpi      = 400
)

Part of Individual Conditional Expectation (ICE) Plot on anyplot.ai.

Other implementations