2.6.1. Introduction to Data Visualization

1. Introduction

Data visualization is the graphical representation of data to help us understand patterns, trends, and insights. In R, data visualization is a core skill for both exploratory data analysis and for communicating results. Good visualizations make complex data accessible and actionable.

2. Why Data Visualization Matters

• Makes data easier to interpret and understand.
• Reveals patterns, trends, and outliers.
• Supports decision-making and communication.
• Essential for both exploratory analysis and explanatory reporting.

3. General Features of Good Plots

• Clearly-labeled axes.
• Text that is large enough to read.
• Axes that are not misleading.
• Data displayed appropriately for its type (categorical, continuous, etc.).
• Clean legends and titles.

4. Exploratory vs. Explanatory Plots

• Exploratory Plots

  • Purpose: Help analysts quickly understand and explore the data.
  • Characteristics:
    • Created rapidly, often with minimal formatting.
    • Used to spot trends, outliers, and data quality issues.
    • Axes and legends are labeled but may not be fully polished.
    • Typically used during the initial phase of analysis.
    • Many are generated as you iterate through different questions.
  • Example:

    # Quick scatter plot to check lab results distribution
    ggplot(labs, aes(x = LBTEST, y = LBORRES)) +
      geom_point()
    

    • Outcome: A basic scatter plot showing lab results by test, useful for spotting patterns or errors.

• Explanatory Plots

  • Purpose: Communicate findings and insights to others (stakeholders, publication, presentation).
  • Characteristics:
    • Carefully designed for clarity, aesthetics, and accuracy.
    • Axes, legends, and titles are descriptive and easy to read.
    • Colors, sizes, and themes are chosen for maximum impact.
    • Only a few are created for each project, but each is highly polished.
    • Used in reports, presentations, and publications.
  • Example:

    # Presentation-ready plot with clear labels and legend
    ggplot(labs, aes(x = LBTEST, y = LBORRES, color = USUBJID)) +
      geom_point(size = 4) +
      labs(
        title = "Lab Results by Test and Subject",
        x = "Lab Test",
        y = "Lab Result (LBORRES)",
        color = "Subject ID"
      ) +
      theme_minimal(base_size = 14) +
      theme(legend.position = "top")
    

    • Outcome: A visually appealing plot, ready for sharing with an audience, with clear differentiation by subject and all elements labeled.

• Summary Table: Exploratory vs. Explanatory

5. Exploratory Plot Example

R Code:

library(ggplot2)
labs <- data.frame(
  USUBJID = rep(c("01-701-101", "01-701-102"), each = 3),
  LBTEST = rep(c("ALT", "AST", "HGB"), 2),
  LBORRES = c(35, 40, 13.5, 38, 32, 14.2)
)
ggplot(labs, aes(x = LBTEST, y = LBORRES)) +
  geom_point()

• Description:
  • Quick scatter plot of lab results by test.
  • Axes are labeled, but plot is not yet polished.

Input Table:

• Expected Outcome:

  

  • A scatter plot showing lab results for each test.

6. Explanatory Plot Example

R Code:

ggplot(labs, aes(x = LBTEST, y = LBORRES, color = USUBJID)) +
  geom_point(size = 4) +
  labs(
    title = "Lab Results by Test and Subject",
    x = "Lab Test",
    y = "Lab Result (LBORRES)",
    color = "Subject ID"
  ) +
  theme_minimal(base_size = 14) +
  theme(legend.position = "top")

• Description:
  • Polished plot with clear labels, larger points, legend, and title.
  • Suitable for presentations or reports.

Expected Outcome:

• A visually appealing plot, easy to interpret, with clear subject differentiation.

**Resource download links**

2.6.1.-Introduction-to-Data-Visualization.zip