Saving, Exporting, and Reusing Plots

Export PMx figures reliably (size, resolution, formats), and build reusable plotting functions and templates.
Tip

Core idea of this lesson: A good figure is not finished until it is exported correctly and can be reproduced consistently.

Learning Objectives

By the end of this lesson, you will be able to:

  • Export plots with controlled size and resolution using ggsave().
  • Choose appropriate file formats for slides, reports, and manuscripts.
  • Understand DPI, width, and height tradeoffs.
  • Build reusable plotting functions for consistent PMx workflows.
  • Separate figure construction from figure export.

Key Ideas

  • A plot is not complete until it is exported correctly for its intended use.
  • Export settings (size, resolution, format) determine how a figure appears outside R.
  • The same plot object can be exported in multiple formats for different purposes.
  • Vector formats (PDF, SVG) preserve quality; raster formats (PNG) depend on resolution.
  • Plot construction and export should be treated as separate steps.
  • Reusable plotting functions ensure consistency across analyses and reports.

Why Exporting Matters

A plot that looks good in the RStudio/Positron viewer can:

  • look blurry in slides,
  • appear cramped in a manuscript,
  • or break layout when resized.

In PMx workflows, figures are often reused across:

  • internal reports,
  • regulatory documents,
  • publications,
  • slide decks.

Export settings must be intentional.

Setup

library(tidyverse)
library(nlme)

data(Remifentanil, package = "nlme")

remi <- as_tibble(Remifentanil) %>%
  rename(
    ID   = ID,
    TIME = Time,
    DV   = conc
  ) %>%
  filter(!is.na(DV))

Worked Example 1: Build a Plot Object

Always assign plots to objects before exporting.

p <- ggplot(remi, aes(TIME, DV, group = ID)) +
  geom_line(alpha = 0.08, color = "grey35") +
  scale_y_log10() +
  labs(
    title = "Remifentanil Concentration–Time Profiles",
    x = "Time (min)",
    y = "Concentration (ng/mL, log10 scale)"
  ) +
  theme_minimal()

p

Worked Example 2: Export with ggsave()

Basic export:

ggsave(
  filename = "remi_profiles.png",
  plot     = p,
  width    = 6,
  height   = 4,
  dpi      = 300
)

What these arguments mean

  • width, height: physical size in inches
  • dpi: resolution (dots per inch)
  • filename extension determines file type
Note

For manuscripts: 300–600 dpi is typical.
For slides: 150–300 dpi is usually sufficient.

File Format Guidance

PNG

  • Good for slides and reports
  • Raster (resolution-dependent)

PDF

  • Vector format
  • Ideal for manuscripts
  • Scales without losing sharpness
ggsave(
  "remi_profiles.pdf",
  plot = p,
  width = 6,
  height = 4
)

SVG

  • Vector format
  • Useful for editing in Illustrator or similar tools

Common Export Problems

  • Saving without specifying size.
  • Resizing inside PowerPoint instead of exporting correctly.
  • Using too low DPI for print.
  • Forgetting to re-export after changing the plot.

Reusable Plot Functions

Instead of rewriting the same code repeatedly, build a function.

plot_profiles <- function(data, log_scale = TRUE) {
  g <- ggplot(data, aes(TIME, DV, group = ID)) +
    geom_line(alpha = 0.08, color = "grey35") +
    labs(
      x = "Time (min)",
      y = "Concentration (ng/mL)"
    ) +
    theme_minimal()
  
  if (log_scale) {
    g <- g + scale_y_log10() +
      labs(y = "Concentration (ng/mL, log10 scale)")
  }
  
  return(g)
}

p2 <- plot_profiles(remi)
p2

Now you have:

  • consistent axis labels
  • consistent styling
  • reproducible figure logic

Separate Construction from Export

Good practice:

p <- plot_profiles(remi)

ggsave("figure1.pdf", p, width = 6, height = 4)
ggsave("figure1.png", p, width = 6, height = 4, dpi = 300)

One plot object → multiple formats.

Multi-Panel Export

If using patchwork or similar layout tools:

library(patchwork)

p_combined <- (p | p)

p_combined

ggsave(
  "remi_combined.pdf",
  p_combined,
  width = 10,
  height = 4
)

Make sure the export size matches the intended layout.

Strategies

  • Always save plot objects.
  • Define export size intentionally.
  • Use PDF for manuscripts, PNG for slides.
  • Keep a consistent width/height across figures in a report.
  • Store plotting functions in a utilities script for reuse.

Common Mistakes

  • Relying on the RStudio/Positron viewer instead of exporting at the final size.
  • Exporting plots before finalizing labels, scales, or themes.
  • Using different sizes or formats for figures within the same report.
  • Hardcoding plot settings repeatedly instead of using reusable functions.
  • Treating export as a final step instead of part of the workflow.
  • Not checking exported figures in their final destination (slides, PDF, etc.).

Practice Problems

  1. Export the same plot at 4×3 inches and 8×6 inches. Compare results.
  2. Export both PNG and PDF versions and inspect sharpness.
  3. Modify plot_profiles() to add a subtitle argument.
  4. Create a function that automatically exports both PNG and PDF.
  5. Decide on a “standard figure size” for your reports and justify it.
export_plot <- function(plot_obj, filename_base,
                        width = 6, height = 4) {
  ggsave(paste0(filename_base, ".pdf"),
         plot_obj, width = width, height = height)
  
  ggsave(paste0(filename_base, ".png"),
         plot_obj, width = width, height = height, dpi = 300)
}

export_plot(p2, "remi_profiles_standard")

Summary

In this lesson, you learned how to:

  • Export figures with intentional size and resolution.
  • Choose appropriate formats (PNG vs PDF).
  • Avoid common export mistakes.
  • Build reusable plotting functions.
  • Separate figure construction from figure delivery.

A PMx figure is complete only when it is:

  • correct,
  • readable,
  • and reproducibly exported.
  • Always assign your plot to an object (e.g., p <- ...) before exporting.
  • Control width and height explicitly — don’t rely on the RStudio/Positron viewer size.
  • Use PDF/SVG when you need vector output (manuscripts, editing) and PNG for slides.
  • For PNGs, start with 300 dpi (higher for print, lower for quick drafts).
  • Export once in the final size you want to appear in your report/slide to avoid blurry resizing.
  • One plot object → export to multiple formats (.pdf + .png) for different deliverables.