Convert reporting_triangle to reviser vintages format
Source:R/as_reviser_vintages.R
as_reviser_vintages.RdThis function converts a reporting_triangle object to a vintages object
in the wide format used by the
reviser package.
reviser provides tools for analysing data revisions and vintages,
including state space nowcasting models and methods for assessing when
releases become stable.
Converting to reviser format enables use of those methods alongside
baselinenowcast's nowcasting functionality.
Arguments
- x
A reporting_triangle object to convert.
- ...
Additional arguments passed to
reviser::vintages_wide().
Value
A tibble of class tbl_pubdate with a time column for the
reference dates and one column per publication date containing the
cumulative reported value as known on that publication date.
Details
A reporting_triangle stores incremental counts at each delay column,
while a reviser vintages object stores the cumulative reported value at
each publication date.
The conversion takes the row-wise cumulative sum of the reporting triangle
and maps each (reference date, delay) cell to a publication date
reference_date + delay in the triangle's delay unit.
The long form is then pivoted to wide format using
reviser::vintages_wide().
NA values in the triangle propagate through the cumulative sum, so any
unobserved cells in the original triangle remain NA in the vintages
output.
To convert back to a reporting_triangle object, use
as_reporting_triangle.tbl_pubdate().
See also
as_reporting_triangle.tbl_pubdate()for converting back
Reporting triangle construction and validation
[.reporting_triangle(),
[<-.reporting_triangle(),
as.data.frame.reporting_triangle(),
as.matrix.reporting_triangle(),
as_ChainLadder_triangle(),
as_reporting_triangle(),
as_reporting_triangle.data.frame(),
as_reporting_triangle.matrix(),
as_reporting_triangle.tbl_pubdate(),
as_reporting_triangle.triangle(),
assert_reporting_triangle(),
get_delays_from_dates(),
get_delays_unit(),
get_max_delay(),
get_mean_delay(),
get_quantile_delay(),
get_reference_dates(),
get_report_dates(),
get_reporting_structure(),
head.reporting_triangle(),
is_reporting_triangle(),
new_reporting_triangle(),
print.reporting_triangle(),
reporting_triangle-class,
summary.reporting_triangle(),
tail.reporting_triangle(),
truncate_to_delay(),
truncate_to_quantile(),
validate_reporting_triangle()
Examples
# Create a reporting triangle from synthetic NSSP data
data_as_of_df <- syn_nssp_df[syn_nssp_df$report_date <= "2026-04-01", ]
rep_tri <- as_reporting_triangle(data = data_as_of_df)
#> ℹ Using max_delay = 154 from data
# Convert to reviser vintages (wide) format
vintages <- as_reviser_vintages(rep_tri)
print(vintages)
#> # Vintages data (publication date format):
#> # Time periods: 159
#> # Vintages: 159
#> time `2025-10-25` `2025-10-26` `2025-10-27` `2025-10-28`
#> <date> <dbl> <dbl> <dbl> <dbl>
#> 1 2025-10-25 194 248 274 287
#> 2 2025-10-26 NA 198 251 280
#> 3 2025-10-27 NA NA 167 225
#> 4 2025-10-28 NA NA NA 145
#> 5 2025-10-29 NA NA NA NA
#> 6 2025-10-30 NA NA NA NA
#> 7 2025-10-31 NA NA NA NA
#> 8 2025-11-01 NA NA NA NA
#> 9 2025-11-02 NA NA NA NA
#> 10 2025-11-03 NA NA NA NA
#> # ℹ 149 more rows
#> # ℹ 155 more variables: `2025-10-29` <dbl>, `2025-10-30` <dbl>,
#> # `2025-10-31` <dbl>, `2025-11-01` <dbl>, `2025-11-02` <dbl>,
#> # `2025-11-03` <dbl>, `2025-11-04` <dbl>, `2025-11-05` <dbl>,
#> # `2025-11-06` <dbl>, `2025-11-07` <dbl>, `2025-11-08` <dbl>,
#> # `2025-11-09` <dbl>, `2025-11-10` <dbl>, `2025-11-11` <dbl>,
#> # `2025-11-12` <dbl>, `2025-11-13` <dbl>, `2025-11-14` <dbl>, …