README.md

In yjunechoe/junebug: A person collection of functions, mostly for interactive use

• junebug Package
  • Installation and usage
• Highlights
  • Themes
  • ggsave_auto() - Accumulate multiple versions of a plot
  • split_track() - Track groups individually over a region of time
  • col_dummy() - Add dummy coded variables as columns

junebug Package

A hodge-podge of functions for various uses, from interactive debugging to preparing data for modeling. The name of the package comes from my nickname and also reflects my early fascination for insects, which has actually contributed to my passion in data visualization.

To some extent, this package was made for personal use, so keep in mind that it’s currently not designed to be light on dependencies.

Installation and usage

devtools::install_github("yjunechoe/junebug")

When using {junebug}, is highly recommended to also load the {tidyverse} set of packages (specifically, {dplyr}, {tibble}, {purrr}, {tidyr}, {ggplot2}, {stringr}).

library(junebug)
# install.packages("tidyverse")
library(tidyverse)

Depending on the function being used, {junebug} may require packages other than the tidyverse packages to also be loaded (details below).

Highlights

Some functions/capabilities of {junebug}

Themes

depends: packages {ggplot2} and {ggtext}

Also need to have some fonts enabled/imported in R ({extrafont} or {showtext} recommended)

theme_junebug_calm

Fonts: Roboto, Roboto Slab

ggplot(iris, aes(Petal.Length, Petal.Width, color = Species)) +
  geom_point(size = 2) +
  theme_junebug_calm() +
  labs(
    title = "Petal profile of plant species in the iris dataset",
    subtitle = "Virginica has the largest petal of the three species"
  ) +
  scale_color_manual(values = junebug_palette("CubanMay", 3)) +
  facet_wrap(~ Species)

theme_junebug_clean

Fonts: Roboto, Roboto Slab

ggplot(iris, aes(Petal.Length, Petal.Width, color = Species)) +
  geom_point(size = 2) +
  theme_junebug_clean() +
  labs(
    title = "Petal profile of plant species in the iris dataset",
    subtitle = "Virginica has the largest petal of the three species"
  ) +
  scale_color_manual(values = junebug_palette("CranberryWhiteGrub", 3)) +
  facet_wrap(~ Species)

theme_junebug_compact

Fonts: Zilla Slab, Futura Bk BT, Roboto Condensed

ggplot(iris, aes(Petal.Length, Petal.Width, color = Species)) +
  geom_point(size = 2) +
  theme_junebug_compact() +
  labs(
    title = "Petal profile of plant species in the iris dataset",
    subtitle = "Virginica has the largest petal of the three species"
  ) +
  scale_color_manual(values = junebug_palette("CranberryWhiteGrub", 3)) +
  facet_wrap(~ Species)

ggsave_auto() - Accumulate multiple versions of a plot

depends: packages {ggplot2} and {magick}

ggsave_auto() is a wrapper around ggplot2::ggsave(). By default, it creates a new folder named “ggsaveauto_img” and saves the last plot printed to “img_001.png”, “img_002.png”, etc. It also has an option (enabled by default) to load the most recently saved plot into view using the system default app.

The follow code saves the plot to “ggsaveauto_img/img_001.png” with width = 20, height = 12, and unit = 'cm'.

ggplot(mtcars, aes(hp, mpg)) +
  geom_point(color = 'red')

ggsave_auto()

Calling ggsave_auto() again after printing a second plot saves the plot to “ggsaveauto_img/img_002.png”

ggplot(mtcars, aes(hp, mpg)) +
  geom_point(color = 'blue')

ggsave_auto()

Names of path and image files can be customized using the following arguments: folder, n_pad, padding, img_prefix, and img_ext.

To turn off the automatic loading of the saved plot, set plot_load = "none". Alternatively, to load the saved plot into view inside RStudio, use plot_load = "rstudio" (but not recommended for quality issues).

split_track() - Track groups individually over a region of time

depends: package {rlang}

split_track() is a helper function for {gganimate}, designed for use with the transition_reveal() engine.

split_track() can be used for a special case of plotting change in a variable (y-axis) over time (x-axis) (e.g., line graphs), where you have multiple groups and want to track them individually over a region.

For example, say we have this following data about four groups (name) whose value changes over time (ranging from 1 to 100).

set.seed(123)
library(gganimate)

plot_df <- tibble(
  a = cumsum(runif(100, -1, 1)),
  b = cumsum(runif(100, -1.5, 1.5)),
  c = cumsum(runif(100, -2, 2)),
  d = cumsum(runif(100, -2.5, 2.5))
) %>% pivot_longer(a:d) %>%
  group_by(name) %>%
  mutate(reveal_time = row_number(), x = row_number()) %>%
  ungroup()

plot_df

With this data, transition_reveal() would animate the trajcetory of all groups simultaneously. If we want to track each group individually over a region (say, where x is greater than or equal to 75), we can do the following:

plot_df %>%
  split_track(grouping_var = name, tracked_along = reveal_time, x >= 75) %>%
  ggplot(aes(x, value, color = name)) +
  geom_line(size = 1) +
  transition_reveal(reveal_time) +
  theme_classic()

You can also specify which groups get tracked individually, using the tracked_groups argument. In the following code, we also change the grouping variable to a factor and specify the ordering of the levels to make sure that the lines that are revealed later are drawn over lines that are revealed earlier.

plot_df %>%
  split_track(grouping_var = name, tracked_along = reveal_time, x >= 75, tracked_groups = c('a', 'c')) %>%
  mutate(name = factor(name, levels = c("b", "d", "a", "c"))) %>% # reordering between groups
  ggplot(aes(x, value, color = name)) +
  geom_line(size = 1) +
  transition_reveal(reveal_time) +
  theme_classic()

col_dummy() - Add dummy coded variables as columns

depends: package {data.table}

Given a column containing n distinct levels, col_dummy() produces n-1 new columns that treatment code each level except the reference level. This explicit coding is necessary for some packages which do not work with implicit coding in R’s factor data type (e.g., the {lavaan} package for structural equation modeling)

levels(iris$Species)

## [1] "setosa"     "versicolor" "virginica"

iris %>% 
  col_dummy(Species, reference = "virginica") %>% 
  head()

## The reference level of 'Species' is 'virginica'

##   Sepal.Length Sepal.Width Petal.Length Petal.Width Speciessetosa
## 1          5.1         3.5          1.4         0.2             1
## 2          4.9         3.0          1.4         0.2             1
## 3          4.7         3.2          1.3         0.2             1
## 4          4.6         3.1          1.5         0.2             1
## 5          5.0         3.6          1.4         0.2             1
## 6          5.4         3.9          1.7         0.4             1
##   Speciesversicolor
## 1                 0
## 2                 0
## 3                 0
## 4                 0
## 5                 0
## 6                 0

Note that the model output is the same:

lm(Sepal.Length ~ Species, data = iris) # implicit coding

## 
## Call:
## lm(formula = Sepal.Length ~ Species, data = iris)
## 
## Coefficients:
##       (Intercept)  Speciesversicolor   Speciesvirginica  
##             5.006              0.930              1.582

lm(Sepal.Length ~ Speciesversicolor + Speciesvirginica, data = col_dummy(iris, Species)) # explicit coding

## The reference level of 'Species' is 'setosa'

## 
## Call:
## lm(formula = Sepal.Length ~ Speciesversicolor + Speciesvirginica, 
##     data = col_dummy(iris, Species))
## 
## Coefficients:
##       (Intercept)  Speciesversicolor   Speciesvirginica  
##             5.006              0.930              1.582

col_dummy() is similar to psych::dummy.code(), except that col_dummy() creates n-1 variables and directly adds the new variables as columns in the data instead of returning a matrix.

yjunechoe/junebug documentation built on July 9, 2021, 7:12 p.m.