Introduction
In grates: Grouped Date Classes

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.align = "center",
  fig.width = 7,
  fig.height = 5
)

Overview

The goal of {grates} is to make it easy to group dates across a range of different time intervals. It defines a collection of classes and associated methods that, together, formalise the concept of grouped dates and are intuitive to use. To assist in formatting plots of grates objects we also provides x-axis scales that can be used in conjunction with {ggplot2} output. Currently implemented classes are:

grates_yearweek;
grates_epiweek;
grates_isoweek;
grates_period;
grates_yearmonth;
grates_yearquarter;
grates_year; and
grates_month

The underlying implementation for these objects build upon ideas of Davis Vaughan and the unreleased {datea} package as well as Zhian Kamvar and the {aweek} package.

grates objects

yearweek, epiweek and isoweek {#yearweek}

yearweek objects are stored as the number of weeks (starting at 0L) from the date of the firstday nearest the Unix Epoch (1970-01-01). Put more simply, the number of seven day periods from:

1969-12-29 for firstday equal to 1 (Monday)
1969-12-30 for firstday equal to 2 (Tuesday)
1969-12-31 for firstday equal to 3 (Wednesday)
1970-01-01 for firstday equal to 4 (Thursday)
1970-01-02 for firstday equal to 5 (Friday)
1970-01-03 for firstday equal to 6 (Saturday)
1970-01-04 for firstday equal to 7 (Sunday)

They can be constructed directly from integers via the new_yearweek() function but it is generally easier to use the either the as_yearweek() coercion function or the yearweek() constructor. as_yearweek() takes two arguments; x, the vector (normally a Date or POSIXt) you wish to group, and firstday, the day of the week you wish your weeks to start on. yearweek() takes three arguments; year and week integer vectors and, again, a firstday value.

The epiweek class is similar to the yearweek class but, by definition, will always begin on a Sunday. They are stored as the integer number of weeks (again starting at 0L) since 1970-01-04 so internally are akin to <grates_yearweek_sunday> objects but with the benefit of slightly more efficient implementations for many of the associated methods.

Likewise, the isoweek class is similar to epiweek class but uses the ISO 8601 definition of a week that will always start on a Monday. Internally they are stored as the integer number of weeks since 1969-12-29.

library(grates)

# Choose some consecutive dates that begin on a Friday
first <- as.Date("2021-01-01")
weekdays(first)
dates <- first + 0:9

# Below we use a Friday-week grouping
weeks <- as_yearweek(dates, firstday = 5L)
(dat <- data.frame(dates, weeks))

# we can also use the constructor function if we already have weeks and years
yearweek(year =c(2020L, 2021L), week = c(1L, 10L), firstday = 5L)

# epiweeks always start on a Sunday
(epiwk <- as_epiweek(Sys.Date()))
weekdays(as.Date(epiwk))

# isoweeks always start on a Sunday
(isowk <- as_isoweek(Sys.Date()))
weekdays(as.Date(isowk))

By default plots (using {ggplot2}) will centre yearweek (epiweek / isoweek) labels:

library(ggplot2)

# use simulated linelist data from the outbreaks package
dat <- outbreaks::ebola_sim_clean
dat <- dat$linelist$date_of_infection

# calculate the total number for across each week
week_dat <- aggregate(
    list(cases = dat),
    by = list(week = as_epiweek(dat)),
    FUN = length
)

head(week_dat)

# plot the output
(week_plot <-
    ggplot(week_dat, aes(week, cases)) + 
    geom_col(width = 1, colour = "white") +
    theme_bw())

We can have non-centred date labels on the x_axis by utilising the associated scale_x_grates functions and explicitly specifying a format for the date labels:

week_plot + scale_x_grates_epiweek(format = "%Y-%m-%d")

Period {#period}

period objects are stored as the integer number, starting at 0L, of periods since the Unix Epoch (1970-01-01) and a specified offset. Here periods are taken to mean groupings of n consecutive days.

Like with yearweek objects, period objects can be constructed directly via a call to new_period() but more easily via the as_period() coercion function. as_period() takes 3 arguments; x, the vector (normally a Date or POSIXt) you wish to group, n, the integer number of days you wish to group, and offset, the value you wish to start counting groups from relative to the Unix Epoch. For convenience, offset can be given as a date you want periods to be relative to (internally this date is converted to integer).

Note that storage and calculation purposes, offset is scaled relative to n. I.e. offset <- offset %% n and values of x stored relative to this scaled offset.

# calculate the total number for across 14 day periods with no offset.
# note - 0L is the default value for the offset but we specify it explicitly
# here for added clarity
period_dat <- aggregate(
    list(cases = dat),
    by = list(period = as_period(dat, n = 14L, offset = 0L)),
    FUN = length
)

head(period_dat)

# lower date bounds are used for the x axis
ggplot(period_dat, aes(period, cases)) +
    geom_col(width = 1, colour = "white") +
    theme_bw( ) +
    theme(axis.text.x = element_text(angle = 45, hjust=1)) +
    xlab("")

# using a date as an offset
start <- as.Date("2020-01-03")
dates <- start + 0:9
offset <- as.Date("2020-01-01")
data.frame(dates, period = as_period(dates, n = 7L, offset = offset))

yearmonth, yearquarter and year {#yearxxx}

yearmonth, yearquarter and year objects are stored as the integer number of months/quarters/years (starting at 0L) since the Unix Epoch (1970-01-01).

Similar to other grates objects we provide both coercion and construction functions.

# calculate the monthly number of cases
(month_dat <- aggregate(
    list(cases = dat),
    by = list(month = as_yearmonth(dat)),
    FUN = length
))

# plot with centred labels
(month_plot <- 
    ggplot(month_dat, aes(month, cases)) + 
    geom_col(width = 1, colour = "white") +
    theme_bw() +
    theme(axis.text.x = element_text(angle = 45, hjust=1)) +
    xlab(""))

# again we can have non-centred date labels by applying the associated scale
month_plot + scale_x_grates_yearmonth(format = "%Y-%m-%d")

# yearquarter works similarly
(quarter_dat <- aggregate(
    list(cases = dat),
    by = list(quarter = as_yearquarter(dat)),
    FUN = length
))

ggplot(quarter_dat, aes(quarter, cases)) + 
    geom_col(width = 1, colour = "white") +
    theme_bw() +
    theme(axis.text.x = element_text(angle = 45, hjust=1)) +
    xlab("")

# year also works similarly
(year_dat <- aggregate(
    list(cases = dat),
    by = list(year = as_year(dat)),
    length
))

ggplot(year_dat, aes(year, cases)) + 
    geom_col(width = 1, colour = "white") +
    theme_bw() +
    theme(axis.text.x = element_text(angle = 45, hjust=1)) +
    xlab("")

# Construction functions can also be used
yearmonth(2022L, 11L)
yearquarter(2022L, 4L)
year(2022L)

month {#month}

month objects are stored as the integer number of n-month groups (starting at 0L) since the Unix Epoch (1970-01-01). Here n-months is taken to mean a 'grouping of n consecutive months'.

<grates_month> objects can be constructed directly from integers via the new_month() function and through coercion via the as_month() function. as_period() takes 2 arguments; x, the vector (normally a Date or POSIXt) you wish to group, n, the integer number of months you wish to group.

# calculate the bimonthly number of cases
(bimonth_dat <- aggregate(
    list(cases = dat),
    by = list(group = as_month(dat, n = 2L)),
    FUN = length
))

# by default lower date bounds are used for the x axis
(bimonth_plot <- 
    ggplot(bimonth_dat, aes(group, cases)) + 
    geom_col(width = 1, colour = "white") +
    theme_bw() +
    theme(axis.text.x = element_text(angle = 45, hjust=1)) +
    xlab(""))

Note that the default plotting behaviour of non-centred date labels is different to that of the yearweek, yearmonth, yearquarter and year scales where labels are centred by default. To obtain centred labels you must explicitly set the format to NULL in the scale:

month_plot + scale_x_grates_yearmonth(format = NULL)

Methods and operations

For all grates objects we have added many methods and operations to ensure logical and consistent behaviour. Where things break down we try to provide detailed messaging explaining why errors have occurred. Whilst this behviour is implemented for all grates objects, below we illustrate how it manifests with epiweek objects.

# use the unique epiweeks from the earlier example
x <- week_dat$week

# min, max and range
(minx <- min(x))
(maxx <- max(x))
(rangex <- range(x))

# seq method works if both `from` and `to` are epiweeks
seq(from = minx, to = maxx, by = 6L)

# but will error informatively if `to` is a different class
try(seq(from = minx, to = 999, by = 6L))

# conversion of yearweek objects back to dates will return the date at the
# lower bound of each yearweek interval
dat <- head(week_dat)
transform(dat, new_date = as.Date(week))

# addition (subtraction) of wholenumbers will add (subtract) the corresponding
# number of weeks to (from) the object
(dat <- transform(dat, plus4 = week + 4L, minus4 = week - 4L))

# addition of two yearweek objects will error as the intention is unclear
try(transform(dat, willerror = week + week))

# Subtraction of two yearweek objects gives the difference in weeks between them
transform(dat, difference = plus4 - minus4)

# epiweeks can be combined with themselves but not other classes (assuming an
# epiweek object is the first entry)
c(minx, maxx)
identical(c(minx, maxx), rangex)
try(c(minx, 1L))