ggplot.decorated: Create a New ggplot for a Decorated Data Frame

In bergsmat/yamlet: Versatile Curation of Table Metadata

Create a New ggplot for a Decorated Data Frame

Description

Creates a new ggplot object for a decorated data.frame. This is the ggplot() method for class 'decorated'. It creates a ggplot object using the default method, but reclassifies it as 'decorated_ggplot' so that a custom print method is invoked; see print.decorated_ggplot.

Usage

## S3 method for class 'decorated'
ggplot(data, ...)

Arguments

data	decorated, see decorate
...	passed to ggplot

Details

This approach is similar to but more flexible than the method for ggready. For fine control, you can switch between 'data.frame' and 'decorated' using as_decorated (supplies null decorations) and as.data.frame (preserves decorations).

Value

return value like ggplot but inheriting 'decorated_ggplot'

See Also

decorate resolve ggready

Other decorated_ggplot: ggplot_build.decorated_ggplot(), print.decorated_ggplot()

Other interface: canonical.decorated(), classified.data.frame(), decorate.character(), decorate.data.frame(), desolve.decorated(), io_csv.character(), io_csv.data.frame(), io_res.character(), io_res.decorated(), io_table.character(), io_table.data.frame(), io_yamlet.character(), io_yamlet.data.frame(), is_parseable.default(), mimic.default(), modify.default(), promote.list(), read_yamlet(), resolve.decorated(), scripted.default(), selected.default(), write_yamlet()

Examples

file <- system.file(package = 'yamlet', 'extdata','quinidine.csv')
library(ggplot2)
library(dplyr)
library(magrittr)
# par(ask = FALSE)

x <- decorate(file)
x %<>% filter(!is.na(conc))

# Manipulate class to switch among ggplot methods.
class(x)
class(data.frame(x))
class(as_decorated(data.frame(x)))

# The bare data.frame gives boring labels and un-ordered groups.
map <- aes(x = time, y = conc, color = Heart)
data.frame(x) %>% ggplot(map) + geom_point()

# Decorated data.frame uses supplied labels.
# Notice CHF levels are still not ordered. (Moderate first.)
x %>% ggplot(map) + geom_point()

# If we resolve Heart, CHF levels are ordered.
x %<>% resolve(Heart)
x %>% ggplot(map) + geom_point()

# We can map aesthetics as decorations.
x %<>% decorate('Heart: [ color: [gold, purple, green]]')
x %>% ggplot(map) + geom_point()

# Colors are matched to particular levels. Purple drops out here:
x %>% filter(Heart != 'Moderate') %>% ggplot(map) + geom_point()

# We can resolve other columns for a chance to enrich the output with units.
x %<>% resolve
suppressWarnings( # because this complains for columns with no units
  x <- modify(x, title = paste0(label, '\n(', units, ')'))
)
x %>% ggplot(map) + geom_point()

# Or something fancier.
x %<>% modify(conc, title = 'conc_serum. (mg*L^-1.)')
x %>% ggplot(map) + geom_point()

# The y-axis title is deliberately given in spork syntax for elegant coercion:
library(spork)
x %<>% modify(conc, expression = as.expression(as_plotmath(as_spork(title))))
x %>% ggplot(map) + geom_point()
# Add a fancier label for Heart, and facet by a factor:
x %<>% modify(Heart, expression = as.expression(as_plotmath(as_spork('CHF^\\*'))))
x %>% ggplot(map) + geom_point() + facet_wrap(~Creatinine)

# ggready handles the units and plotmath implicitly for a 'standard' display:
x %>% ggready %>% ggplot(map) + geom_point() + facet_wrap(~Creatinine)

# Notice that instead of over-writing the label
# attribute, we are creating a stack of label
# substitutes (title, expression) so that
# label is still available as an argument
# if we want to try something else.  The
# print method by default looks for all of these.
# Precedence is expression, title, label, column name.
# Precedence can be controlled using
# options(decorated_ggplot_search = c(a, b, ...) ).

# Here we try a dataset with conditional labels and units.

file <- system.file(package = 'yamlet', 'extdata','phenobarb.csv')
x <- file %>% decorate %>% resolve
# Note that value has two elements for label and guide.
x %>% decorations(value)

# The print method defaults to the first, with warning.
map <- aes(x = time, y = value, color = event)

x %>% ggplot(map) + geom_point()


# If we subset appropriately, the relevant value is substituted.
x %>% filter(event == 'conc') %>% ggplot(map) + geom_point()

x %>% filter(event == 'conc') %>%
ggplot(aes(x = time, y = value, color = ApgarInd)) + geom_point()

x %>% filter(event == 'dose') %>%
ggplot(aes(x = time, y = value, color = Wt)) +
geom_point() +
scale_y_log10() +
scale_color_gradientn(colours = rainbow(4))

# print.decorated_ggplot will attempt to honor coordinated aesthetics.
x <- data.frame(x = c(1:6, 3:8), y = c(1:6,1:6), z = letters[c(1:6,1:6)])
x %<>% decorate('z: [color: ["red", "blue", "green", "gold", "black", "magenta"]]')
x %<>% decorate('z: [fill: ["red", "blue", "green", "gold", "black", "magenta"]]')
x %<>% decorate('z: [shape: [20, 21, 22, 23, 24, 25]]')
x %<>% decorate('z: [linetype: [6, 5, 4, 3, 2, 1]]')
x %<>% decorate('z: [alpha: [ .9, .8, .7, .6, .5, .4]]')
x %<>% decorate('z: [size: [1, 1.5, 2, 2.5, 3, 3.5]]')
x %>% ggplot(aes(
 x, y,
  color = z,
  fill = z,
  shape = z,
  linetype = z, 
  alpha = z,
  size = z,
)) + 
  geom_point() +
  geom_line(size = 1)

bergsmat/yamlet documentation built on Feb. 18, 2024, 5:50 a.m.