Visualizing Data
Zhaoxia Yu
ggplot
- ggplot is based on grammar of graphics.
- It is a powerful and flexible system for creating a wide range of visualizations.
Step 1: Data
Step 2: Map to Aesthetics
Step 3: Add Geometric Layer
Additional Features
Additional Features
Code
# label diagnosis 0, 1, 2 to healthy, impaired, and Alzheimer's
alzheimer_data <- alzheimer_data %>%
mutate(diagnosis_name = factor(diagnosis,
levels = c(0, 1, 2),
labels = c("Healthy", "Impaired", "Alzheimer's")))
ggplot(data = alzheimer_data,
aes(x = diagnosis_name)) +
geom_bar() +
labs(
title = "Distribution of Diagnosis",
x = "Diagnosis Type",
y = "Count"
) +
theme_minimal()
Additional Features
Categorical by Categorical Variable
- Stacked Bar Plot (Counts)
Categorical by Categorical Variable
- Grouped (Dodged) Bar Plot (Counts)
Code
ggplot(alzheimer_data, aes(x = factor(female, labels = c("Male", "Female")),
fill = factor(diagnosis, labels = c("Healthy", "Impaired", "Alzheimer's")))) +
geom_bar(position = "dodge") +
labs(
title = "Count of Diagnosis by Gender (Grouped)",
x = "Gender",
y = "Count",
fill = "Diagnosis"
) +
theme_minimal()
Categorical by Categorical Variable
- Proportions (Within Groups)
Code
ggplot(alzheimer_data, aes(x = factor(female, labels = c("Male", "Female")),
fill = factor(diagnosis, labels = c("Healthy", "Impaired", "Alzheimer's")))) +
geom_bar(position = "fill") + # Normalizes to 100% per group
labs(
title = "Proportion of Diagnosis by Gender",
x = "Gender",
y = "Proportion",
fill = "Diagnosis"
) +
scale_y_continuous(labels = scales::percent) + # Show as %
theme_minimal()
Categorical by Categorical Variable
- Standardized Proportions (Comparable Across Groups)
Code
library(dplyr)
prop_data <- alzheimer_data %>%
count(female, diagnosis) %>%
group_by(female) %>%
mutate(prop = n / sum(n) * 1000) # Standardize per 1000
ggplot(prop_data, aes(x = factor(female, labels = c("Male", "Female")),
y = prop,
fill = factor(diagnosis, labels = c("Healthy", "Impaired", "Alzheimer's")))) +
geom_col(position = "dodge") +
labs(
title = "Standardized Diagnosis Rates by Gender (per 1000)",
x = "Gender",
y = "Cases per 1000",
fill = "Diagnosis"
) +
theme_minimal()
Visualize a Numerical Variable: boxplot
- The volume of left hippocampus (lhippo) in the alzheimer_data dataset is a numerical variable.
Visualize a Numerical Variable: histogram
Visualize a Numerical Variable by a Categorical Variable
Code
ggplot(data = alzheimer_data,
mapping = aes(
x = factor(female, labels = c("Male", "Female")), # Convert to labeled factor
y = lhippo,
fill = factor(female, labels = c("Male", "Female")) # Optional: color by gender
)) +
geom_boxplot() +
labs(
title = "Boxplot of Hippocampal Volume by Gender",
x = "Gender",
y = "Left Hippocampal Volume (cm³)",
fill = "Gender" # Legend title
) +
theme_minimal()
Visualize a Numerical Variable by a Categorical Variable
Code
ggplot(data = alzheimer_data,
mapping = aes(
x = lhippo,
fill = factor(female, labels = c("Male", "Female")) # Ensure 'female' is a labeled factor
)) +
geom_histogram(bins = 35, alpha = 0.7, position = "identity") + # Adjust transparency & position
labs(
title = "Histogram of Left Hippocampal Volume by Gender",
x = "Left Hippocampal Volume (cm3)",
y = "Count",
fill = "Gender" # Legend title
) +
theme_minimal() +
scale_fill_manual(values = c("Male" = "skyblue", "Female" = "salmon")) # Custom colors (optional)
Summary
