Create signal object — to_signal • rplanes

This function creates an object of the S3 class "signal". The user can conditionally specify either a "forecast" or "observed" signal.

Usage

to_signal(
  input,
  outcome,
  type = "observed",
  resolution = "weeks",
  horizon = NULL
)

Arguments

input: Data to be converted to signal; see "Details" for more information
outcome: Name of the outcome column in the input data
type: Signal type; must be one of "observed" or "forecast"; default is "observed"
resolution: The temporal resolution of the signal; data can be aggregated daily, weekly, or monthly; default is "weeks"; see "Details" for more information
horizon: Number of time steps ahead for forecast signals; only used if type="forecast"; default is NULL

Value

An object of the class signal. The object will have a second class of either observed or forecast depending on the value passed to the "type" argument.

Details

The input signal data may be either "observed" or "forecast" data. Depending on the type, the input data must conform to certain format prior to submission. In both cases, the data must be passed as a data frame.

For "observed" data the data frame must at minimum include columns for location (geographic unit such as FIPS code) and date (date of reported value; must be date class). The data should also include a column that contains the outcome (e.g., case count).

For "forecast" data the data frame must include columns for location (geographic unit such as FIPS code), date (date corresponding to forecast horizon; must be date class or character formatted as 'YYYY-MM-DD'), horizon (forecast horizon), lower (the lower limit of the prediction interval for the forecast), point (the point estimate for the forecast), and upper (the upper limit of the prediction interval for the forecast). Note that the read_forecast function returns data in this format.

The input data must at the daily, weekly, or monthly resolution. The "resolution" parameter is designed to use string matching. This allows flexibility for the user, such that, for example, an input of "day", "days", or "daily" would all resolve to a resolution of days. The same rules apply for designating weekly or monthly resolution.

Examples

hosp <- read.csv(system.file("extdata/observed/hdgov_hosp_weekly.csv", package = "rplanes"))
to_signal(hosp, outcome = "flu.admits", type = "observed", resolution = "weeks")

fp <- system.file("extdata/forecast/2022-10-31-SigSci-TSENS.csv", package = "rplanes")
ex_forecast <- read_forecast(fp)
to_signal(ex_forecast, outcome = "flu.admits", type = "forecast", horizon = 4, resolution = "weeks")