Date and time scales: scale_date

Mix.install([
  {:giggly, path: Path.join(__DIR__, "..")},
  {:kino, "~> 0.19"}
  ])

alias Giggly.Plot

render = fn plot ->
  plot
  |> Giggly.to_svg()
  |> IO.iodata_to_binary()
  |> Kino.Image.new(:svg)
end

About this notebook

This notebook follows the examples on ggplot2's reference page for scale_x_date(), which covers scale_y_date(), scale_x_datetime(), scale_y_datetime(), scale_x_time() and scale_y_time() too, in the order the page gives them. Each example shows ggplot2's code first and then Giggly's. An example Giggly cannot draw yet keeps its ggplot2 code and names what it is waiting for, so the page is also a list of what remains.

These are the position scales for dates, date-times and times of day, which ggplot2 adds by itself for a column of one of them. Setting one by hand chooses its limits, where its breaks fall and how its labels are written, in units of time such as "1 week" and formats such as "%b %d". Giggly has Giggly.Plot.scale_x_date/2 and Giggly.Plot.scale_y_date/2, with their limits, breaks, labels and minor breaks, breaks also as a step of time (date_breaks: [week: 1]) and labels as a Calendar.strftime/3 format (date_labels: "%b %d"); it has no scale_x_datetime() or scale_x_time(), since date-times and times of day have no scale yet. A plot is described with Giggly.Plot and drawn with Giggly.to_svg/2, which returns an SVG document. The setup cell aliases Giggly.Plot, so the cells write Plot.new/2, and defines render, which draws a plot and shows it below the cell with Kino.Image.

Breaks and labels

Sys.Date() is today's date, so the data is the last 30 days, each with a random price.

last_month <- Sys.Date() - 0:29
set.seed(1)
df <- data.frame(
  date = last_month,
  price = runif(30)
)
base <- ggplot(df, aes(date, price)) +
  geom_line()

# The date scale will attempt to pick sensible defaults for
# major and minor tick marks. Override with date_breaks, date_labels
# date_minor_breaks arguments.
base + scale_x_date(date_labels = "%b %d")
base + scale_x_date(date_breaks = "1 week", date_labels = "%W")
base + scale_x_date(date_minor_breaks = "1 day")

Giggly draws the first and third plots. The cell takes today's date from the clock, as Sys.Date() does, though Date.utc_today/0 gives the date in UTC where R gives the local one, and writes out the 30 prices R's runif() draws after set.seed(1). A step of time is written as Date.shift/2 takes one, [day: 1] for "1 day", and a format is Calendar.strftime/3's, which writes "%b %d" as R does, Oct 19.

today = Date.utc_today()

df = %{
  "date" => Enum.map(0..29, fn days -> Date.add(today, -days) end),
  "price" => [
    0.2655086631421, 0.37212389963679016, 0.5728533633518964, 0.9082077899947762,
    0.2016819310374558, 0.8983896849676967, 0.9446752686053514, 0.6607977924868464,
    0.6291140438988805, 0.061786270467564464, 0.20597457489930093, 0.17655675252899528,
    0.6870228466577828, 0.384103718213737, 0.7698414199985564, 0.49769924208521843,
    0.7176185082644224, 0.9919060948304832, 0.3800351794343442, 0.7774452213197947,
    0.9347052311059088, 0.21214252128265798, 0.651673766085878, 0.12555509596131742,
    0.26722066872753203, 0.3861140925437212, 0.013390333158895373, 0.3823879570700228,
    0.86969084572047, 0.3403489966876805
  ]
}

base =
  df
  |> Plot.new(%{x: :date, y: :price})
  |> Plot.geom_line()

base
|> Plot.scale_x_date(date_labels: "%b %d")
|> render.()

base
|> Plot.scale_x_date(date_minor_breaks: [day: 1])
|> render.()

Not in Giggly yet: "%W" in date_labels, the week of the year, which Calendar.strftime/3 cannot write.

Limits

# Set limits
base + scale_x_date(limits = c(Sys.Date() - 7, NA))

R's NA leaves the upper limit to the data, today; Giggly writes it nil. The line through the days before the limit is not drawn.

base
|> Plot.scale_x_date(limits: {Date.add(today, -7), nil})
|> render.()