Describing and Visualising Data

Overview

This chapter introduces fundamental data visualisation techniques using the grammar of graphics approach through ggplot2. We’ll use real ecological data to create and customise informative plots while building on the statistical concepts from the previous chapter.

Data Introduction

We’ll work with two datasets from the Long Term Ecological Research (LTER) Network:

Fiddler Crabs (pie_crab): Body size measurements of fiddler crabs, collected to study population characteristics.
Sugar Maples (hbr_maples): Tree growth measurements under different environmental conditions.

Grammar of Graphics with ggplot2

The ggplot2 package uses a layered approach to “paint a canvas” one component at a time:

Data: The dataset we’re visualising
Aesthetics: How variables map to visual elements (what goes on which axis)
Geometries: The type of plot that best shows your data patterns
Additional layers: Labels, themes, and customisation for clear communication

Building Plots Step by Step

Just like writing field methods, build your plots systematically:

Start with data (ggplot(data))
Map variables (aes(x = var1, y = var2))
Add plot type (geom_*())
Enhance with labels and themes

Core plot types

Histograms: to understand distributions

Histograms help us see patterns in measurements, like how animal sizes or plant heights are distributed in a population.

# Step 1: Start with our data and map size to x-axis
ggplot(pie_crab, aes(x = size)) +
  # Step 2: Create histogram with 30 bins for detailed view
  # More bins = more detail, but might show random variation
  # Fewer bins = smoother pattern, but might miss important details
  geom_histogram(bins = 30) +
  # Step 3: Add clear labels for scientific communication
  labs(
    x = "Carapace Width (mm)",
    y = "Count",
    title = "Distribution of Fiddler Crab Sizes"
  )

# Add mean line
ggplot(pie_crab, aes(x = size)) +
  geom_histogram(bins = 30) +
  geom_vline(aes(xintercept = mean(size, na.rm = TRUE)),
    color = "red", linetype = "dashed"
  ) +
  labs(
    x = "Carapace Width (mm)",
    y = "Count",
    title = "Fiddler Crab Sizes with Mean"
  )

Boxplots: Comparing Groups

Think of a boxplot like a summary of crab sizes at each beach site - just as you might summarize different areas in an ecological survey. Each box tells a story about the crabs living there.

Let’s build this plot step by step:

# First, let's make a simple comparison between sites
# Just like comparing different study areas in the field
ggplot(pie_crab, aes(x = site, y = size)) +
  geom_boxplot() +
  labs(
    x = "Site",
    y = "Carapace Width (mm)",
    title = "Fiddler Crab Sizes by Site"
  )

# Now let's make it easier to see differences
# by adding colors - like marking different study zones
ggplot(pie_crab, aes(x = site, y = size, fill = site)) +
  geom_boxplot() +
  labs(
    x = "Site",
    y = "Carapace Width (mm)",
    title = "Fiddler Crab Sizes by Site"
  )

The box in the middle shows where most of your crabs are (like finding the most common sizes in your sample). The line in the middle is the typical size for that site. This helps us quickly see if crabs are generally larger at some sites than others - just like you might find different sized plants in different parts of a forest.


### Bar Plots: Counting and Comparing

Bar plots help us answer questions like "How many samples did we collect?" or "What's the typical size at each site?" - common questions in field surveys.

First, let's count how many crabs we found at each site:

::: {.cell}

```{.r .cell-code}
# Count crabs at each site (like tallying field observations)
ggplot(pie_crab, aes(x = site)) +
  geom_bar() +  # geom_bar automatically counts for us
  labs(
    x = "Site",
    y = "Number of Crabs",
    title = "Number of Crabs per Site"
  )

:::

Now, let’s look at the typical size of crabs at each site. Just like you might measure trees or plants in different areas to see where they grow best:

# Calculate average size for each site
# Think of this like summarizing field measurements
pie_crab %>%
  # Organize data by site
  group_by(site) %>%
  # Calculate the average and its reliability
  summarise(
    mean_width = mean(size, na.rm = TRUE), # average size
    se = sd(size, na.rm = TRUE) / sqrt(n()) # how reliable is our average?
  ) %>%
  # Create the plot
  ggplot(aes(x = site, y = mean_width)) +
  geom_col() + # make the bars
  # Add error bars to show reliability
  geom_errorbar(
    aes(
      ymin = mean_width - se,
      ymax = mean_width + se
    ),
    width = 0.2
  ) +
  labs(
    x = "Site",
    y = "Average Carapace Width (mm)",
    title = "Typical Crab Size at Each Site"
  )

The height of each bar shows the average size, and the black lines (error bars) tell us how confident we are about that average - just like you might be more confident about your findings when you have more samples.


### Scatterplots: Exploring Relationships

Examine maple growth patterns:

::: {.cell}

```{.r .cell-code}
# Basic scatterplot
ggplot(hbr_maples, aes(x = year, y = stem_length)) +
  geom_point() +
  labs(
    x = "Year",
    y = "Stem Length (cm)",
    title = "Sugar Maple Growth Over Time"
  )

# Add trend line and group by treatment
ggplot(hbr_maples, aes(
  x = year, y = stem_length,
  color = watershed
)) +
  geom_point() +
  geom_smooth(method = "lm") +
  labs(
    x = "Year",
    y = "Stem Length (cm)",
    title = "Sugar Maple Growth by Watershed"
  )

`geom_smooth()` using formula = 'y ~ x'

:::

Essential Plot Modifications

Clear Labels and Titles

ggplot(pie_crab, aes(x = site, y = size)) +
  geom_boxplot() +
  labs(
    x = "Collection Site",
    y = "Carapace Width (mm)",
    title = "Fiddler Crab Size Distribution",
    subtitle = "Comparing sizes across collection sites",
    caption = "Data: PIE LTER"
  )

Appropriate Colours

# Using colour-blind friendly palette
ggplot(pie_crab, aes(
  x = site, y = size,
  fill = site
)) +
  geom_boxplot() +
  scale_fill_brewer(palette = "Set2") +
  labs(
    x = "Collection Site",
    y = "Carapace Width (mm)",
    title = "Fiddler Crab Sizes by Site (with colour)"
  )

Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set2 is 8
Returning the palette you asked for with that many colors

Scale Adjustments

# Adjust axis limits and breaks
ggplot(hbr_maples, aes(
  x = year, y = stem_length,
  color = watershed
)) +
  geom_point() +
  scale_y_continuous(
    limits = c(0, max(hbr_maples$stem_length, na.rm = TRUE)),
    breaks = seq(0, max(hbr_maples$stem_length, na.rm = TRUE), by = 50)
  ) +
  labs(
    x = "Year",
    y = "Stem Length (cm)"
  )

Theme Customisation

# Using minimal theme with customisation
ggplot(pie_crab, aes(x = site, y = size)) +
  geom_boxplot() +
  labs(
    x = "Collection Site",
    y = "Carapace Width (mm)",
    title = "Fiddler Crab Size Distribution"
  ) +
  theme_minimal() +
  theme(
    axis.text = element_text(size = 10),
    axis.title = element_text(size = 12),
    plot.title = element_text(size = 14, face = "bold"),
    legend.position = "bottom"
  )