Plotting in R

This section focuses on R-based visualizations for building science and indoor environmental quality data.

The examples in this library will mostly use:

Core ideas

1. Data as data frames

Most R plotting in this library assumes:

  • your data is in a data frame or tibble
  • one row = one observation
  • one column = one variable

If your data is not in this shape, reshaping is usually the first step using functions such as:

  • pivot_longer() and pivot_wider() (tidyr)
  • mutate(), filter(), group_by(), summarise() (dplyr)

Well-structured data makes plotting simpler and more predictable.

2. Grammar of Graphics mindset

ggplot2 follows a “grammar of graphics” approach where:

  • data provides values
  • aesthetics map variables to visual properties
  • geoms define how data is drawn (points, lines, bars, etc.)
  • scales control how data maps to visual values
  • facets create small multiples
  • themes control appearance

Most plots follow a structure like:

ggplot(data, aes(x = <var>, y = <var>, color = <group>)) +
  geom_<something>() +
  theme_<something>() +
  labs(x = "X label", y = "Y label", title = "Title")

This makes plots systematic, testable, and easy to refine.

3. Factor ordering matters

For categorical variables (e.g. survey responses, building IDs, Likert scales):

  • set factor levels explicitly
  • never rely on default alphabetical ordering

Example:

df$response <- factor(
  df$response,
  levels = c("Cold", "Cool", "Neutral", "Warm", "Hot")
)

This avoids confusing axis and legend orderings.

4. Themes and consistency

Visual consistency matters as much as correctness.

You should aim for:

  • a consistent font family and size
  • restrained use of color
  • minimal distractions (no unnecessary gridlines)
  • consistent sizing across figures

For now, theme_minimal() or theme_classic() are good defaults. Later, a custom lab theme can be added and reused everywhere.

5. Exporting figures

Export figures with explicit size and resolution:

ggsave(
  "figures/example_plot.pdf",
  width = 8,
  height = 5,
  units = "in",
  dpi = 300
)

Recommendations:

  • use vector formats (PDF, SVG) for publications
  • embed intent into filenames
  • keep figures in a dedicated folder

6. Reproducibility principles

Plots should be:

  • generated from code, not GUIs
  • fully reproducible from raw data
  • styled using functions, not copy-paste

Aim to write small helper functions such as:

  • theme_lab()
  • scale_color_lab()
  • scale_fill_lab()

and reuse them across projects.

Where this R section is going

This space will grow to include:

  • reusable plotting patterns
  • helper functions
  • common domain patterns
  • references and recommended reading

The goal is a growing knowledge base, not just a gallery.