fill_missing: Fill missing data in a data set to prepare it for use within...

In epiforecasts/EpiNow2: Estimate Real-Time Case Counts and Time-Varying Epidemiological Parameters

Fill missing data in a data set to prepare it for use within the package

Description

This function ensures that all days between the first and last date in the data are present. It adds an accumulate column that indicates whether modelled observations should be accumulated onto a later data point. point. This is useful for modelling data that is reported less frequently than daily, e.g. weekly incidence data, as well as other reporting artifacts such as delayed weekedn reporting. The function can also be used to fill in missing observations with zeros.

Usage

fill_missing(
  data,
  missing_dates = c("ignore", "accumulate", "zero"),
  missing_obs = c("ignore", "accumulate", "zero"),
  initial_accumulate,
  obs_column = "confirm",
  by = NULL
)

Arguments

data	Data frame with a date column. The other columns depend on the model that the data are to be used, e.g. estimate_infections() or estimate_secondary(). See the documentation there for the expected format. The data must not already have an accumulate function, otherwise the function will fail with an error.
missing_dates	Character. Options are "ignore" (the default), "accumulate" and "zero". This determines how missing dates in the data are interpreted. If set to "ignore", any missing dates in the observation data will be interpreted as missing and skipped in the likelihood. If set to "accumulate", modelled observations on dates that are missing in the data will be accumulated and added to the next non-missing data point. This can be used to model incidence data that is reported less frequently than daily. In that case, the first data point is not included in the likelihood (unless initial_accumulate is set to a value greater than one) but used only to reset modelled observations to zero. If "zero" then all observations on missing dates will be assumed to be of value 0.
missing_obs	Character. How to process dates that exist in the data but have observations with NA values. The options available are the same ones as for the missing_dates argument.
initial_accumulate	Integer. The number of initial dates to accumulate if missing_dates or missing_obs is set to "accumulate". This argument needs ot have a minimum of 1. If it is set to 1 then no accumulation is happening on the first data point. If it is greater than 1 then dates are added to the beginning of the data set to get be able to have a sufficient number of modelled observations accumulated onto the first data point. For modelling weekly incidence data this should be set to 7. If accumulating and the first data point is not NA and this is argument is not set, then that data point will be removed with a warning.
obs_column	Character (default: "confirm"). If given, only the column specified here will be used for checking missingness. This is useful if using a data set that has multiple columns of hwich one of them corresponds to observations that are to be processed here.
by	Character vector. Name(s) of any additional column(s) where missing data should be processed separately for each value in the column. This is useful when using data representing e.g. multiple geographies. If NULL (default) no such grouping is done.

Value

a data.table with an accumulate column that indicates whether values are accumulated (see the documentation of the data argument in estimate_infections())

Examples

cases <- data.table::copy(example_confirmed)
## calculate weekly sum
cases[, confirm := data.table::frollsum(confirm, 7)]
## limit to dates once a week
cases <- cases[seq(7, nrow(cases), 7)]
## set the second observation to missing
cases[2, confirm := NA]
## fill missing data
fill_missing(cases, missing_dates = "accumulate", initial_accumulate = 7)

epiforecasts/EpiNow2 documentation built on Dec. 10, 2024, 5:32 p.m.