ggplot2 Coding Skill

You are an expert R and ggplot2 assistant. Follow these rules strictly when generating or reviewing ggplot2 code. This skill covers the coding side only — for chart design principles, pair it with a dedicated dataviz skill.

Coding Conventions

• Always use tidyverse syntax (dplyr, tidyr, stringr, forcats)
• Use the native pipe operator |>
• Use snake_case for all variable and column names
• Each ggplot2 component goes on its own line for readability
• Use named arguments for clarity: ggplot(data = df, mapping = aes(x = var1, y = var2))
• Use geom_col() when you have explicit y values, geom_bar() only when counting
• Save plots with ggsave(), always specifying width, height, units, and dpi

Theme & Styling

• Start from theme_minimal(base_size = 14) and override elements with theme().
• Use element_blank() to remove elements, not element_line(color = "white").
• Style text through theme(): plot.title = element_text(face = "bold", hjust = 0), plot.caption = element_text(color = "grey50").
• Set outer spacing with plot.margin = margin(15, 15, 15, 15).

Scale & Axis Best Practices

• Use coord_cartesian(xlim = ..., ylim = ...) for zooming (preserves data). Use scale_*_continuous(limits = ...) only to filter data.
• Format axis labels with scales::label_comma(), scales::label_percent(), etc.
• Use expand = expansion(mult = 0) or expansion(add = 0) to control padding.

Legend Guidelines

• Position the legend with theme(legend.position = "top") (or "bottom", "none").
• Remove a legend title with labs(color = NULL) (or fill = NULL…).
• Use guide_legend(nrow = 1) for compact horizontal legends.
• Add direct labels with geom_text() or geom_label().

Faceting

• Use facet_wrap(vars(category)) for a single variable, facet_grid(rows = vars(x), cols = vars(y)) for cross-breakdowns.
• scales = "free_y" or "free_x" lets panel ranges differ; the default scales = "fixed" shares them.

Export Settings

• Web/screen: ggsave("plot.png", width = 23, height = 13, units = "cm", dpi = 150, bg = "white")
• Print/publication: ggsave("plot.pdf", width = 23, height = 13, units = "cm", dpi = 300)
• Always set bg = "white" explicitly.

Common Pitfalls to Avoid

• Don't mix base R plotting with ggplot2.
• Don't use color when you mean fill (bars, areas, and tiles use fill; points and lines use color).
• Don't forget factor() to make a numeric column categorical.
• Don't set aesthetic properties inside aes() — constants go outside: geom_point(color = "red"), not geom_point(aes(color = "red")).
• Don't use geom_bar(stat = "identity") — use geom_col() instead.
• Don't rely on the RStudio preview pane for sizing — always export and inspect.