In sarahhbellum/NobBS: Nowcasting by Bayesian Smoothing

NobBS v. 1.1.0

Updates from v1.0.0 include: Quantile estimates - NobBS now returns quantile estimates to support quantile-based evaluation metrics such as the Weighted Interval Score (WIS). Batched reports - Added functionality to handle batched reporting of cases. Day-of-the-week effect - Incorporated adjustments to account for systematic variations in reporting by day of the week. Three new vignettes have been added to demonstrate the usage of these features. The package maintainer has changed from Sarah McGough to Rami Yaari.

NobBS v. 1.0.0

Updates to v. 0.1.0 (Initial Release) are major fixes to remove deprecated tidyverse functions, replacing group_by_ with group_by(!!sym()) and vars_select with standard select.

NobBS v. 0.1.0 (Initial Release)

See README.md for a full tutorial on using the NobBS package.

Updates to the development version include:

• A function for stratified Bayesian nowcasting, NobBS.strat
• An argument to specify the proportion_reported and inflate the estimated cases by that factor, important in outbreaks with severe under-reporting or asymptomatic cases.
• Additional checks and warnings issued for impossible temporal combinations of moving window sizes, maximum delay D, and the length of the time series.
• Warning issued when a case onset week is skipped in the time series or is missing reports altogether.
• Plot examples

NobBS v. 0.0.0.9000 (Development)

NobBS (Nowcasting by Bayesian Smoothing) is a new R package that produces smooth Bayesian nowcasts of incomplete, time-stamped reporting data implemented from McGough et al. 2019.

sarahhbellum/NobBS documentation built on June 12, 2025, 10:57 a.m.