A Plotting Assistant for Fast ggplot2 Visualisations • plot2

The Single-Line Plotting Assistant for `ggplot2`

The ggplot2 R package composes any plot from layers, aesthetics, scales, coordinates and themes. This grammar makes it flexible and precise, but it also means that creating a plot requires the knowledge of many functions and arguments: a plot of “counts per department, coloured by year” requires knowledge of geom_col(), aes(fill = ), position_dodge(), scale_fill_*() and theme(), and of the data steps that come before them.

plot2 takes the opposite approach. Like base R’s plot(), one call gives a complete plot, even when the type of plot or other details are not given. You describe what you want to see, and plot2() summarises, sorts and labels the data, and chooses the plot type, colours, axes and theme:

iris |> plot2()
#> Using x = Sepal.Length
#> Using y = Sepal.Width
#> Using category = Species
#> Using type = "point" since both axes are numeric

iris |> plot2(x = Species)
#> Using y = Sepal.Length
#> Using type = "boxplot" since all groups in Species contain at least three values

iris |> plot2(x = Species, y = mean(Sepal.Length))

The result is a standard ggplot object, so every ggplot2 function and extension still works on it.

Installation

plot2 is not on CRAN yet. Install it from r-universe:

install.packages("plot2", repos = c("https://msberends.r-universe.dev", "https://cloud.r-project.org"))

What plot2 replaces

Task	ggplot2	plot2
Summarise, then plot	`group_by()` + `summarise()` + `ggplot()` + `geom_col()`	`plot2(x = a, y = mean(b))`
Sort bars	`mutate(fct_reorder(...))`	`x.sort = "freq-desc"`
Group by colour	`aes(fill = var)` for bars, `aes(colour = var)` for points, each with its own `scale_*()`	`category = var`
Set colours	`scale_fill_manual(values = ...)` or `scale_colour_manual(values = ...)`, depending on the geom	`colour = c(...)`
Facet	`+ facet_wrap(~var)`	`facet = var`
Data labels	`+ geom_text(aes(label = ...), vjust = ...)`	automatic for columns
Stacked bars	`geom_col(position = "stack")` or `position = "fill"`	`stacked = TRUE` or `stacked_fill = TRUE`

Plot types

plot2() supports all ggplot2 geoms, and adds plot types such as dumbbell, back-to-back, Sankey, spider and UpSet plots. Each of the 20 plots below is made with only a single line of code:

Grid of 24 example plots made with plot2(), including bar, line, area, boxplot, violin, heatmap, Sankey, spider, UpSet, back-to-back and map plots

Show the code for each plot

# Column
admitted_patients |> plot2(hospital, n())

# Bar
admitted_patients |> plot2(hospital, n(), ward, type = "bar")

# Bar with percentages
ggplot2::mpg |> plot2(manufacturer, type = "barpercent")

# Stacked to 100%
admitted_patients |> plot2(age_group, n(), hospital, stacked_fill = TRUE)

# Line
admitted_patients |> plot2(format(date, "%Y"), n(), ward, type = "line")

# Line with points
admitted_patients |> plot2(format(date, "%Y"), n(), ward, type = "linedot")

# Area
admitted_patients |> plot2(format(date, "%Y"), n(), hospital, type = "area", stacked = TRUE, x.complete = 0)

# Point
iris |> plot2()

# Beeswarm
iris |> plot2(Species, Sepal.Length, type = "beeswarm")

# Boxplot
admitted_patients |> plot2(hospital, age, gender)

# Violin
admitted_patients |> plot2(ward, age, type = "violin")

# Histogram
admitted_patients |> plot2(age, type = "histogram")

# Density
iris |> plot2(Sepal.Length, category = Species, type = "density", colour_transparency = 0.5)

# Dumbbell
admitted_patients |> plot2(hospital, median(age), gender, type = "dumbbell")

# Heatmap
mtcars |> cor() |> plot2(colour = c("#2166AC", "white", "#B2182B"))

# Sankey
Titanic |> plot2(c(Class, Age, Survived), category = Sex, type = "sankey")

# Spider
admitted_patients |> plot2(age_group, n(), hospital, type = "spider")

# UpSet
mtcars |> transform(manual = am == 1, straight = vs == 1, gears_4 = gear == 4, mpg_20 = mpg > 20) |> plot2(x = c(manual, straight, gears_4, mpg_20), type = "upset")

# Back-to-back
admitted_patients |> plot2(age_group, n(), facet = gender, type = "back-to-back")

# Map
netherlands |> plot2(datalabels = FALSE)

Examples

The four plotting directions x, y, category and facet describe most plots. They accept columns and calculations:

admitted_patients |>
  plot2(x = age_group,
        y = n(),
        category = gender,
        facet = ward)
#> Using free y scales per facet since the two y scales differ by 25% or more (use facet.fixed_y = TRUE for the same scale in all facets)

admitted_patients |>
  plot2(x = age_group,
        y = n_distinct(patient_id),
        category = case_when(gender == "F" ~ "Female", gender == "M" ~ "Male"),
        facet = paste("Ward:", ward),
        
        # additional options are each only 1 argument:
        x.title = "Age Group",
        y.title = "Unique Nr of Patients",
        category.title = "Gender",
        facet.fixed_y = TRUE)
#> Assuming facet.repeat_labels_y = FALSE since y has fixed scales

NOTE

All arguments are named after the plotting direction they apply to: x.*, y.*, category.* and facet.*. With plot2(), there is no need to remember argument or function names, or to look them up in the documentation. Type the direction and press Tab, and your editor lists every option for it, together with its description:

Titles can be calculated from the data as well:

mtcars |>
  plot2(x = mpg,
        y = hp,
        category = factor(cyl),
        category.title = "Cylinders",
        title = paste("Horsepower of", n(), "cars"))
#> Using type = "point" since both axes are numeric

A correlation matrix becomes a heatmap:

mtcars |>
  cor() |>
  plot2(datalabels = TRUE,
        colour = c("steelblue", "white", "firebrick"),
        category.title = "Correlation")
#> Assuming type = "tile" since the matrix contains identical row and column names
#> Using category.midpoint = 0 (the current category scale centre)

Learning more

Introduction to plot2 describes the main ideas.
Plot types shows every plot type.
Colour and themes describes how to set manual/organisational style
Fonts explains how any font can be used directly.
Maps shows how maps can be plotted easily.

create_interactively() opens a Shiny app that builds a plot from a point-and-click interface and shows the code:

iris |> create_interactively()

Contributing

Issues, feature requests and pull requests are welcome at https://github.com/msberends/plot2.

Licence

GPL-2. See the LICENSE file for details.

plot() + ggplot2 = plot2()

The Single-Line Plotting Assistant for ggplot2