Aggregate and Pad Signals for Reporting
aggregate_pad_signals.RdThis function combines the final weekly signal results into aggregate counts (e.g. total cases, any alarms, number of alarms), and conditionally pads the time series with historical expected values and thresholds prior to the signal generation window. This is primarily used for report generation and visualization.
Arguments
- signal_results
A tibble returned by [get_signals()], containing weekly signal detection results (cases, alarms, upperbound, expected, etc.).
- preprocessed
A data frame containing the surveillance data preprocessed with [preprocess_data()].
- number_of_weeks
Integer specifying how many weeks to include in the aggregation.
- method
A character string specifying the method used to generate the signals. Determines whether padding is necessary. For `"glm"` methods, padding is skipped as it is assumed to be already included.
Value
A named list with two elements:
- signals_agg
A tibble with aggregated results per stratum, including total cases, whether any alarms occurred, and the number of alarms in the last `number_of_weeks`.
- signals_padded
A tibble with the original `signal_results` augmented with additional rows containing historical `expected` and `upperbound` values (if padding was applied).
Details
Padding is applied only for non-GLM methods. It reconstructs the expected trajectory before the signal detection window, which is useful for plotting full time series trends.
Examples
if (FALSE) { # \dontrun{
results <- get_signals(preprocessed_data, method = "farrington")
output <- aggregate_pad_signals(results, number_of_weeks = 6, method = "farrington")
output$signals_agg
output$signals_padded
} # }