analyse_modobs2: Analyse modelled values versus observed data.
In stineb/rbeni: Useful R functions for Beni.

analyse_modobs2

R Documentation

Analyse modelled values versus observed data.

Description

Calculates a set of performance statistics and optionally creates plots of modelled versus observed values.

Usage

analyse_modobs2(
  df,
  mod,
  obs,
  type = "points",
  filnam = NA,
  relative = FALSE,
  xlim = NULL,
  ylim = NULL,
  use_factor = NULL,
  shortsubtitle = FALSE,
  rsquared = TRUE,
  plot_subtitle = TRUE,
  plot_linmod = TRUE,
  plot_legend = TRUE,
  label = FALSE,
  id = NULL,
  nlabels = 1,
  ...
)

Arguments

`df`	A data frame containing columns with names corresponding to arguments `mod` and `obs`
`mod`	A character string specifying the variable name (column) of the modelled (simulated) values in data frame `df`.
`obs`	A character string specifying the variable name (column) of the observed values in data frame `df`.
`type`	If `"points"`, uses `geom_points()`, if `"hex"` uses `ggplot2::geom_hex()`, if `"heat"` uses adjusted `geom_points()` with color indicating density, if `"density"` uses `stat_density_2d()` to draws polygos of equal density.
`filnam`	A character string specifying the name of the file containing the plot. Defaults to `NA` (no file is created).
`relative`	A logical specifying whether the relative RMSE and bias (after division by the mean) is to be showed in the subtitle labels.
`shortsubtitle`	A boolean specifying whether to display a reduced set of metrics in the subtitle.
`rsquared`	A boolean specifying whether to display R-squared and the RMSE (if `TRUE`) or the r (Pearson's correlation coefficient) and the p (p-value of test of significance of correlation, if `TRUE`). Defaluts to `TRUE`.
`plot_subtitle`	A boolean specifying whether to display any metrics. Defaults to `TRUE`.
`plot_linmod`	A boolean specifying whether to display the fitted linear regression as a red line. Defaults to `TRUE`.
`plot_legend`	A boolean specifying whether to display a legend for the colors. Defaults to `TRUE` if `type` is one of `"heat"`, `"hex"`, or `"density"`.
`label`	A boolean specifying whether points should be labelled using ggrepel. Defaults to `FALSE`. Only available for `type == "points"`. Use argument `nlabels` to specify how many points should be labelled, starting with points that have the largest residuals from the linear regression fit.
`id`	A character string specifying the column name that identifies the points. The column's values must be of type integer and is used to label points in case of `label = TRUE`.
`nlabels`	An integer specifying how many points to be labelled, starting with points that have the largest residuals from the linear regression fit. Only available for `type == "points"`. Defaults to one.