求助：如何在R的ggplot2中为面板数据叠加多组密度图

First, ensure you have the required libraries loaded:

library(prodest)
library(tidyverse) # Includes dplyr, tidyr, ggplot2

Step 1: Reshape Data to Long Format #

ggplot works best with long-format data. Convert your wide data (with separate Y and Yhat columns) into long format where we have a column indicating the variable type (Y/Yhat) and the corresponding value:

sim_data_long <- sim_data %>%
  pivot_longer(
    cols = c(Y, Yhat),
    names_to = "variable",
    values_to = "value"
  )

Option 1: Faceted Plots (Recommended for Many ID Variables) #

If you have a large number of idvar levels (like 100 in your example), faceting by idvar keeps the visualization clean. Each facet shows the density of Y (gray) and Yhat (unique color for that ID):

ggplot(sim_data_long, aes(x = value)) +
  # Plot Y density in gray for all IDs
  geom_density(
    data = filter(sim_data_long, variable == "Y"),
    color = "gray50",
    linewidth = 0.8
  ) +
  # Plot Yhat density with unique color per ID
  geom_density(
    data = filter(sim_data_long, variable == "Yhat"),
    aes(color = factor(idvar)),
    linewidth = 0.8
  ) +
  # Create a facet for each ID
  facet_wrap(~idvar) +
  # Customize labels and theme
  labs(
    x = "Value",
    y = "Density",
    color = "ID Variable"
  ) +
  theme_minimal() +
  # Hide legend since each facet corresponds to one ID
  theme(legend.position = "none")

Option 2: Single Plot (For Fewer ID Variables) #

If you have fewer idvar levels, you can overlay all densities on a single plot. Y densities are gray, and each Yhat density uses a unique color:

ggplot() +
  # Plot all Y densities in gray (semi-transparent for overlap visibility)
  geom_density(
    data = sim_data,
    aes(x = Y),
    color = "gray50",
    alpha = 0.3
  ) +
  # Plot each Yhat density with unique color by ID
  geom_density(
    data = sim_data,
    aes(x = Yhat, color = factor(idvar)),
    alpha = 0.7
  ) +
  labs(
    x = "Value",
    y = "Density",
    color = "ID Variable"
  ) +
  theme_minimal()

Key Notes #

• Color Mapping: In both options, Y densities are kept consistent (gray) to focus attention on the shifted Yhat distributions.
• Line Width/Alpha: Adjust linewidth and alpha parameters to improve readability when multiple lines overlap.
• Legend Handling: For faceted plots, the legend is redundant since each facet represents one ID, so we hide it.

内容的提问来源于stack exchange，提问作者user27808