Process raw epidemiological data for the GP model

Validates and prepares input data into the shape consumed by infer_kernel_params() and gp_predict(): it completes the site-by-time grid, drops sites that cannot be modelled, assigns a factor site id, and returns the observations and coordinates as separate, ready-to-use frames.

Usage

data_process(data, ..., drop_zero = FALSE)

Arguments

data

A data frame containing site identifiers, time t, counts n, and coordinates lat and lon.

...

Bare (unquoted) column names that jointly identify a site, e.g. data_process(raw, region, facility_name).

drop_zero

Logical; passed to data_missing() – also drop sites whose observed counts sum to zero (default FALSE).

Value

A list with three elements ready to pass to the model functions:

obs_data

Observations with the site-identifier columns, the factor id, time t, and the count column y_obs (NA where missing).

coordinates

One row per site with id, lon and lat.

nt

The number of time points.