在R语言中展示单变量分布可使用哪些Geom函数？

Hey there! Let's break down your questions about visualizing univariate distributions with ggplot2's Geom functions in R. I've got you covered with practical examples and clear explanations.

Univariate distribution visualization in R relies heavily on the ggplot2 package, where Geom functions act as the building blocks for your plots. The core workflow is straightforward:

1. Load the ggplot2 package.
2. Initialize a ggplot object with your dataset, and map your single variable to the x (or y) aesthetic.
3. Add the appropriate geom_*() layer to render the distribution.

Here's a concrete example using the built-in mtcars dataset to visualize the distribution of miles per gallon (mpg):

# Load the ggplot2 package first
library(ggplot2)

# Create a histogram for mpg distribution
ggplot(mtcars, aes(x = mpg)) +
  geom_histogram(binwidth = 2, fill = "steelblue", color = "white") +
  labs(title = "Distribution of Miles Per Gallon", 
       x = "MPG", 
       y = "Number of Vehicles")

In this code:

• aes(x = mpg) tells ggplot we're focusing on the single variable mpg.
• geom_histogram() creates the histogram, with binwidth controlling how wide each bar is, and fill/color adjusting styling.
• labs() adds descriptive labels to make the plot easier to interpret.

There are several Geom functions tailored to different aspects of univariate distributions. Here are the most commonly used ones:

• geom_histogram(): The go-to for showing frequency/count distribution of continuous data. Great for spotting peaks, gaps, and skewness. Adjust bin size with binwidth to balance detail and readability.
• geom_density(): Renders a smooth kernel density curve, which highlights the overall shape of the distribution (e.g., normal, skewed) without relying on binning. Use alpha to add transparency if overlaying with other elements:

  ggplot(mtcars, aes(x = mpg)) +
    geom_density(fill = "coral", alpha = 0.5)
  

• geom_freqpoly(): A line-based alternative to histograms, plotting the frequency of each bin as a polygon. Useful for comparing multiple distributions side-by-side (works just as well for single variables too).
• geom_boxplot(): Summarizes distribution using quartiles, median, and outliers. Perfect for quickly identifying spread and extreme values. For single variables, map to the y aesthetic and leave x blank:

  ggplot(mtcars, aes(x = "", y = mpg)) +
    geom_boxplot(width = 0.5) +
    labs(x = "") # Remove empty x-axis label
  

• geom_violin(): Combines the best of boxplots and density curves, showing both statistical summaries and the full shape of the distribution. Often paired with a boxplot for clarity:

  ggplot(mtcars, aes(x = "", y = mpg)) +
    geom_violin(fill = "purple", alpha = 0.3) +
    geom_boxplot(width = 0.1, color = "black")
  

• geom_dotplot(): Uses stacked dots to represent frequency, ideal for small datasets where you want to see individual data points. Adjust binwidth to control how dots are grouped.
• geom_jitter(): Adds random "jitter" to points along one axis, preventing overlap when plotting raw data. Useful for visualizing the density of observations in a single variable:

  ggplot(mtcars, aes(x = mpg, y = 1)) +
    geom_jitter(width = 0.5, height = 0) +
    labs(y = "") # Remove unnecessary y-axis label
  

内容的提问来源于stack exchange，提问作者Roger Nebin