Nowcast right-truncated surveillance data

Estimate final counts for recent, incomplete weeks using baselinenowcast. The last max_delay weeks are replaced; earlier observations are unchanged. Downward revisions are redistributed with baselinenowcast::preprocess_negative_values().

Usage

get_ncast(x, max_delay = 2, draws = 1000, prop_delay = 0.5, scale_factor = 3)

Arguments

x

An insightcast_data object with revision history.

max_delay

Number of recent weeks to nowcast. Defaults to 2.

draws

Number of posterior samples. Defaults to 1000.

prop_delay

Proportion of reference times used to estimate delays. Must be between 0 and 1. Defaults to 0.5.

scale_factor

Multiplier applied to max_delay to set the estimation window. Defaults to 3.

Value

An insightcast_ncast object with the shared backbone (key, target, window, interval, history) plus:

data

Corrected series with ncast_lower and ncast_upper 95% credible interval bounds.

meta

Nowcast settings and a per-series ncast_summary.

Details

Weekly data only; other cadences are rejected (the rest of the pipeline is cadence-agnostic).

Author

Cyril Geismar

Examples

if (FALSE) { # \dontrun{
x <- get_data(pathogen = "covid", geo_value = "ca", revisions = TRUE)
ncast <- get_ncast(x)
autoplot(ncast)
} # }