R语言中如何实现条形图叠加散点图？par(fig)绘图叠加异常问题求助

Hey there! Let's work through fixing your overlaid scatter plot and bar plot issue in R — it's a common hurdle for new users, so don't worry about missing details earlier.

First, let's break down what was going wrong in your original code:

• The fig parameter follows the order c(x_left, x_right, y_bottom, y_top) (all values range from 0 to 1, where 0 is left/bottom and 1 is right/top). Your original fig=c(1,0,...) reversed the x-axis bounds, which threw off the entire plot positioning.
• Calling dev.off() right after creating the histogram closed the graphics device, meaning all your subsequent par() settings weren't applied to the plot you were actually trying to build.
• The space=5 in barplot() added massive gaps between bars, making the density plot look disconnected and misaligned.
• Margin (mar) settings and new=TRUE placement weren't tuned to prevent overlap between the two plots.

Here's the corrected code, with comments explaining each key change:

# Generate your sample data
y1 <- rnorm(1000, mean=0, sd=1)
y2 <- rnorm(1000, mean=0, sd=1)

# Calculate histogram density data without plotting immediately
# Using plot=FALSE avoids creating an extra plot we don't need
y_hist <- hist(y1, freq=FALSE, breaks=100, plot=FALSE)

# Set up the scatter plot first (occupies the bottom 75% of the figure)
# mar=c(bottom, left, top, right) adjusts margins to avoid overlap
par(fig=c(0, 1, 0, 0.75), mar=c(4, 4, 1, 1), font.axis=2, cex.axis=1.1)
plot(y1, y2, pch=21, bg="red", xlim=c(-3,3), ylim=c(-3,3),
     xlab="Hodnoty y1", ylab="Hodnoty y2") # Added axis labels for clarity

# Now overlay the bar plot on the top 25% of the figure
# new=TRUE tells R to add to the existing device instead of creating a new one
par(fig=c(0, 1, 0.75, 1), mar=c(0, 4, 3, 1), new=TRUE)
barplot(y_hist$density, axes=FALSE, horiz=FALSE, ylim=c(0,0.4), space=0, ann=FALSE)
# Add a y-axis for the density plot (optional but helpful for readability)
axis(side=2, font.axis=2, cex.axis=1.1)
mtext("Hustota", side=2, line=2.5, font=2)

# Add title and subtitle at the top of the figure
title("Hustota normálního rozdělení", side=3, line=1, font.main=2, cex.main=1.1)
mtext("(generováno 1000 čísel s parametry mean=0, sd=1)", side=3, line=0, font.sub=2, cex.sub=1.1)

Key Fixes Explained: #

• Skipped dev.off(): Instead of plotting the histogram right away, we used plot=FALSE to just capture the density data. This keeps the graphics device open for our combined plot.
• Corrected fig coordinates: The scatter plot uses c(0,1,0,0.75) (spans full width, bottom 0 to 0.75), and the bar plot uses c(0,1,0.75,1) (full width, top 0.75 to 1) — no reversed axes, so plots stack correctly.
• Tweaked margins: The scatter plot has a small top margin, and the bar plot has no bottom margin, so they sit flush without overlapping.
• Reduced bar spacing: space=0 makes bars adjacent, matching the histogram's bin layout so the density plot makes visual sense.
• Fixed subtitle details: Adjusted the parameter description to match your rnorm() call (you had sigma=0, -1 but used mean=0, sd=1).

This should give you a clean, non-overlapping plot where the scatter plot sits below and the density bar plot sits neatly on top.

内容的提问来源于stack exchange，提问作者Hana Liškařová