Densify a reporting triangle into per-series reference x delay count matrices

Densify a reporting triangle into per-series reference x delay count matrices

Usage

reporting_triangle_matrix(
  triangle,
  max_delay,
  value_col = attr(triangle, "value_col")
)

Arguments

triangle

A `csfmt_reporting_triangle_v3`.

max_delay

Number of delay columns (delay 0 .. max_delay-1, in weeks).

value_col

Which value column to reshape (default the triangle's `value_col`; pass a denominator column to reshape that instead).

Value

Named list (by time_series_id) of `list(reference, mat)`, where `mat` is a reference x delay count matrix (zeros filled within the observed region).

See also

Neither package vignette covers this function. It is the densification step every nowcast engine runs first. Reach for it directly only when you want the raw reference x delay matrix rather than an ensemble.

Other reporting triangle functions: csfmt_reporting_triangle_v3()

Examples

w <- cstime::dates_by_isoyearweek$isoyearweek
i <- match("2023-01", w)
set.seed(1)
d <- data.table::data.table(
  isoyearweek_reference = w[i + rep(0:39, each = 3)],
  isoyearweek_reporting = w[i + rep(0:39, each = 3) + rep(0:2, 40)],
  numerator = rpois(120, c(30, 15, 5)),
  indicator_tag = "x", location_code = "nation", age = "total", sex = "total"
)
d <- d[isoyearweek_reporting <= w[i + 39]]
tri <- csfmt_reporting_triangle_v3(
  d,
  id_cols = c("indicator_tag", "location_code", "age", "sex")
)

m <- reporting_triangle_matrix(tri, max_delay = 3)
names(m)
#> [1] "d8da72e3fbb5fd29"

# rows are reference weeks, columns are delays 0, 1, 2
head(m[[1]]$reference, 3)
#> [1] "2023-01" "2023-02" "2023-03"
head(m[[1]]$mat, 3)
#>       0  1 2
#> [1,] 26 20 8
#> [2,] 38 16 3
#> [3,] 24 17 6

# the newest weeks are only partly reported: the later delays are still zero
tail(m[[1]]$mat, 3)
#>        0  1 2
#> [38,] 27 21 3
#> [39,] 24 20 0
#> [40,] 20  0 0