Summary  
This chapter covers how to implement code to create and customize histograms for distribution analysis by dividing numeric data into bins, adjusting bin widths, and interpreting shapes such as skewness and modality.

General domain of usage  
Exploratory data analysis

Histograms are a **fundamental tool** in data analysis for visualizing the distribution of numeric variables. By dividing the range of data into intervals, or **"bins,"** and counting how many data points fall into each bin, a histogram provides a **visual summary** of how values are spread, clustered, or dispersed. This makes it easier to spot patterns, outliers, and the overall shape of your data, which is crucial before applying statistical models or drawing conclusions.

library(ggplot2)

# Sample data: heights of individuals
heights <- c(160, 165, 170, 172, 168, 175, 180, 178, 182, 169, 174, 177, 185, 172, 167)

# Basic histogram
ggplot(data.frame(heights), aes(x = heights)) +
  geom_histogram(binwidth = 5, fill = "skyblue", color = "black") +
  labs(title = "Histogram of Heights",
       x = "Height (cm)",
       y = "Count")

# Adjusting bin width for more or fewer bins
ggplot(data.frame(heights), aes(x = heights)) +
  geom_histogram(binwidth = 2, fill = "orange", color = "black") +
  labs(title = "Histogram of Heights (Smaller Bin Width)",
       x = "Height (cm)",
       y = "Count")

When you examine a histogram, the shape of the bars tells you about the underlying distribution of your data. If the histogram is roughly symmetrical and bell-shaped, your data may be **normally distributed**. If the bars stretch further to the right, with a long tail, the distribution is **right-skewed** (`positively skewed`); if the tail is to the left, it is **left-skewed** (`negatively skewed`). You may also observe **modality**:
- A single prominent peak (`unimodal`);
- Two peaks (`bimodal`);
- More than two peaks, indicating clusters or subgroups in your data.

Interpreting these shapes helps you decide on further analysis steps and informs you about potential data issues or interesting patterns.

library(testthat)

# Source user code and capture the last ggplot expression
.user_env <- new.env(parent = globalenv())
source("user_code.R", local = .user_env)

# Extract the ggplot object: either an assigned variable or the last expression
.plot_obj <- NULL
for (nm in ls(.user_env)) {
    obj <- get(nm, envir = .user_env)
    if (inherits(obj, "ggplot")) {
        .plot_obj <- obj
        break
    }
}
# If not assigned to a variable, re-evaluate the last expression to capture it
if (is.null(.plot_obj)) {
    .exprs <- parse("user_code.R")
    .last_expr <- .exprs[[length(.exprs)]]
    .plot_obj <- eval(.last_expr, envir = .user_env)
}

temperatures <- get("temperatures", envir = .user_env)

# --- CASE 1: temperatures variable ---
test_that("temperatures variable exists, is numeric, and has correct length", {
    expect_true(exists("temperatures", envir = .user_env),
        info = "Variable 'temperatures' should exist.")
    expect_true(is.numeric(temperatures),
        info = "'temperatures' should be numeric.")
    expect_equal(length(temperatures), 20,
        info = "'temperatures' should have 20 values.")
})

# --- CASE 2: ggplot object ---
test_that("histogram plot is created with correct binwidth and labels", {
    expect_true(inherits(.plot_obj, "ggplot"),
        info = "A ggplot histogram should be created.")

    # Check for geom_histogram layer
    has_hist <- any(sapply(.plot_obj$layers, function(layer) {
        inherits(layer$geom, "GeomBar")
    }))
    expect_true(has_hist, info = "Plot should contain a histogram layer.")

    # Check binwidth = 2
    has_binwidth <- any(sapply(.plot_obj$layers, function(layer) {
        bw <- layer$stat_params$binwidth
        !is.null(bw) && bw == 2
    }))
    expect_true(has_binwidth, info = "Histogram should use binwidth = 2.")

    # Check labels
    labs_list <- .plot_obj$labels
    expect_true(!is.null(labs_list$title) && nzchar(labs_list$title),
        info = "Plot should have a title.")
    expect_true(!is.null(labs_list$x) && nzchar(labs_list$x),
        info = "Plot should have an x-axis label.")
    expect_true(!is.null(labs_list$y) && nzchar(labs_list$y),
        info = "Plot should have a y-axis label.")
})


test_main.R

Master the art of data visualization in R using the powerful ggplot2 package. This course guides you from the basics of plotting to advanced visualization techniques, enabling you to create insightful and visually appealing graphics for data analysis and presentation.

Comprehensive exploration of information visualization in R using ggplot2, from foundational concepts to advanced techniques.

Histograms for Distribution Analysis

Ratkaisu