如何用ggplot创建带单X轴的双Y轴分面网格并复刻SigmaPlot图形

Alright, let's figure out how to replicate that SigmaPlot 4x4 grid in ggplot—including those empty first two slots in the first column for your legend! Here's a practical, step-by-step solution tailored to your data setup:

First, let's build the core plot with your mappings: Station on X, value on Y, shape for quarter, color for species, and faceted by constituent in a 4-column grid. I'll add some theme tweaks to align closer to SigmaPlot's clean style:

library(ggplot2)
library(patchwork) # For flexible grid layout (way easier than gridExtra here)
library(dplyr)
library(purrr)

# Replace `your_data` with your actual dataset name
base_plot <- ggplot(your_data, aes(x = Station, y = value)) +
  geom_point(aes(shape = quarter, color = species), size = 3) + # Make points easy to distinguish
  facet_wrap(~constituent, ncol = 4) + # 4 columns to get the 4x4 grid
  # Customize scales for better visibility
  scale_shape_manual(values = c(1, 2, 5, 6)) # Adjust based on how many levels `quarter` has
  scale_color_brewer(palette = "Set2") # Use a distinct color palette for species
  # Theme adjustments to match SigmaPlot's vibe
  theme_bw() +
  theme(
    panel.grid.major = element_line(color = "gray90"),
    panel.grid.minor = element_blank(),
    strip.background = element_rect(fill = "gray85"),
    strip.text = element_text(size = 10, face = "bold"),
    axis.text = element_text(size = 9),
    axis.title = element_text(size = 11, face = "bold")
  )

We need to pull the legend out of the base plot so we can place it in those empty first-column slots:

# Extract the legend (using cowplot's helper function—install with `install.packages("cowplot")` if needed)
library(cowplot)
plot_legend <- get_legend(base_plot)

# Remove the legend from the original plot
plot_no_legend <- base_plot + theme(legend.position = "none")

The key part is creating a 4x4 layout where the first two slots in column 1 are merged to hold the legend, and the rest hold your faceted plots. We'll use patchwork's custom layout syntax for this:

# Split your faceted plot into a list of individual panels
plot_panels <- plot_no_legend$facet$plot_env$data %>%
  distinct(constituent) %>%
  pull(constituent) %>%
  map(~ plot_no_legend %+% filter(your_data, constituent == .x))

# Define the 4x4 layout: "A" is the merged legend slot, letters B-O are your panels
custom_layout <- "
AABC
AADE
FGHI
JKLM
"

# Combine everything into the final grid
final_plot <- (plot_legend + 
                plot_panels[[1]] + plot_panels[[2]] + plot_panels[[3]] +
                plot_panels[[4]] + plot_panels[[5]] + plot_panels[[6]] +
                plot_panels[[7]] + plot_panels[[8]] + plot_panels[[9]] + plot_panels[[10]] +
                plot_panels[[11]] + plot_panels[[12]] + plot_panels[[13]] + plot_panels[[14]]) +
  plot_layout(design = custom_layout) +
  plot_annotation(theme = theme(plot.margin = margin(10, 10, 10, 10)))

# View the final plot
final_plot

Quick Adjustments for Edge Cases: #

• If you have exactly 16 constituent levels, replace two of the plot_panels entries with empty panels (use ggplot() + theme_void() for blank slots) in positions 1 and 5 (first column, rows 1 and 2), then place the legend in one of those slots.
• Tweak the scale_shape_manual and scale_color_brewer values to match the exact shapes/colors from your SigmaPlot figure.

内容的提问来源于stack exchange，提问作者Anonymous coward