insight.cast is an R package for infectious disease nowcasting and forecasting. It was developed through Insight Net, a CDC Center for Forecasting and Outbreak Analytics initiative.
Use it to fetch or validate data, correct reporting delays, compare models and produce forecasts.
Installation
Install the development version from GitHub:
# install.packages("pak")
pak::pak("ACCIDDA/insight.cast")Quick start
library(insight.cast)
tail(example_data)
#> # A tibble: 6 × 5
#> as_of location target target_end_date observation
#> <date> <chr> <chr> <date> <dbl>
#> 1 2025-12-07 CA wk inc flu hosp 2025-12-06 233
#> 2 2025-12-14 CA wk inc flu hosp 2025-12-06 259
#> 3 2025-12-07 NY wk inc flu hosp 2025-12-06 1160
#> 4 2025-12-14 NY wk inc flu hosp 2025-12-06 1171
#> 5 2025-12-14 CA wk inc flu hosp 2025-12-13 412
#> 6 2025-12-14 NY wk inc flu hosp 2025-12-13 1462
fcast <- example_data |>
check_data() |>
get_ncast() |>
get_cv(eval_start_date = as.Date("2024-10-01")) |>
get_fcast()
fcast
#> <insightcast_fcast>
#> Target: wk inc flu hosp
#> Series: 2 (location)
#> Forecast: 2025-12-20 to 2026-01-10 (h = 4)
#> Models: 4 + ENSEMBLE
fcast |> autoplot()
Save a forecast for myRespiLens:
Citation
To cite insight.cast:
citation("insight.cast")
#> To cite package 'insight.cast' in publications use:
#>
#> Geismar C (2026). _insight.cast: Tools for Epidemic Forecasting_. R package
#> version 0.0.1, <https://github.com/ACCIDDA/insight.cast>.
#>
#> A BibTeX entry for LaTeX users is
#>
#> @Manual{,
#> title = {insight.cast: Tools for Epidemic Forecasting},
#> author = {Cyril Geismar},
#> year = {2026},
#> note = {R package version 0.0.1},
#> url = {https://github.com/ACCIDDA/insight.cast},
#> }Acknowledgements
insight.cast uses baselinenowcast and fable. It returns forecasts in hubverse format for submission to CDC forecast hubs.
