Interpolate data gaps — run_interp • SondePolishR

Uses the specified method to fill NA values for a given parameter within the dataset.

Usage

run_interp(data_interp, y_var, method, freq = 1)

data_interp: data.frame from prep_interp() with dupes condensed.
y_var: Variable being interpolated.
method: The method to use for interpolation, options include: linear, spline, random_forest, and ts_interp see details for further information about the methods.
freq: The period interval for the time series in days, only used if method is ts_interp.

a data_interp with a extra column added:

Currently the follow interpolation options are supported:

linear: Uses the zoo::na.approx() function to linearly fill gaps.
spline: Uses the zoo::na.spline() function to fill gaps via spline interpolation (makes smoother curves).
random_forest: Depending on how much data is in the project, this option can be a little slow. Uses missForest::missForest() to fill gaps. uses responses across all the available parameters to try and fill in the missing parameter.
ts_interp: Uses forecast::na.interp() which treats the data as a time series and make predictions based on the natural seasonal patterns within the data. The seasonality is determined by freq. For instance, if set to 1, it will account for daily fluctuations. If set to 365 it will look at annual fluctuations.

interp_dfs <- prep_interp(example_sondeproj)
filled_yvar <- run_interp(interp_dfs$interp, "fDOM_QSU", "linear")