In reichlab/cdcForecastUtils: Utility functions for CDC forecasts

knitr::opts_chunk$set(echo = TRUE, message=FALSE, warning=FALSE)

This vignette explains how to verify submission files for the CDC COVID-19 ILI challenge. Submission files must follow specific formatting guidelines.

There are 2 ways to verify submission files using the cdcForecastUtils package. First, if you have not installed the cdcForecastUtils package, then please do so using devtools:

devtools::install_github("reichlab/cdcForecastUtils")

When this package is installed, you can load it in to your current session:

library(cdcForecastUtils)

Then, choose how to verify using the methods in the examples below.

Verify with verify_entry_file()

This method verifies using a file path without saving the forecasts in the R environment.

# For national/Regional-level forecasts
weekly_forecast_filepath1 <- "/directory/EW10-2020-regional_forecast_file.csv"
verify_entry_file(weekly_forecast_filepath1)

# For state-level forecasts
weekly_forecast_filepath2 <- "/directory/EW10-2020-state_forecast_file.csv"
verify_entry_file(weekly_forecast_filepath2, challenge = "state_ili")

Verify with verify_entry()

This method is almost the same as the previous method, except the forecasts are read into R environment with read_entry(), which the user can inspect if the verification were to fail.

# For national/Regional-level forecasts
weekly_forecast_file1 <- read_entry("/directory/EW10-2020-regional_forecast_file.csv")
verify_entry(weekly_forecast_file1)

# For state-level forecasts
weekly_forecast_file2 <- read_entry("/directory/EW10-2020-state_forecast_file.csv")
verify_entry(weekly_forecast_file2)

For both methods, there will be no errors if the forecast file is formatted correctly. Please refer to the template for more details regarding formatting. Note that there could be warnings or/and messages for missing locations or/and missing targets, which can be ignored if the forecasts for those targets or/and locations are intentionally not included.

reichlab/cdcForecastUtils documentation built on May 6, 2020, 10:43 a.m.