Skip to contents

Plot population overlay goodness-of-fit (GOF) plots

Source: R/plot_gof.R
plot_gof.Rd

Creates a population overlay plot showing central tendency lines for observed (DV), population predicted (PRED), and individual predicted (IPRED) values. Colors and aesthetics for each variable are controlled through the style argument via style_gof(). Use the shown argument to selectively hide variables.

Usage

plot_gof(
  data,
  dv_var = "DV",
  pred_var = "PRED",
  ipred_var = "IPRED",
  time_var = "TIME",
  ntime_var = "NTIME",
  shown = NULL,
  id_var = NULL,
  dose_var = "DOSE",
  loq = NULL,
  loq_method = 0,
  blq_mode = c("obs", "all"),
  cent = c("mean", "mean_sdl", "mean_sdl_upper", "median", "median_iqr", "none"),
  dosenorm = FALSE,
  ref = NULL,
  log_y = FALSE,
  show_caption = TRUE,
  style = NULL
)

Arguments

data

Input dataset.

dv_var

Column containing the dependent variable (DV). Accepts bare names or strings. Default is DV.

pred_var

Column containing population predictions (PRED). Accepts bare names or strings. Default is PRED.

ipred_var

Column containing individual predictions (IPRED). Accepts bare names or strings. Default is IPRED.

time_var

Column containing the actual time variable. Accepts bare names or strings. Default is TIME.

ntime_var

Column containing the nominal time variable. Accepts bare names or strings. Default is NTIME.

shown

Layer visibility settings created by plot_gof_shown(). Defaults can be viewed by running plot_gof_shown() with no arguments.

id_var

Column to group observations for spaghetti lines. Accepts bare names or strings. Default is NULL (no spaghetti lines).

dose_var

Column to use in dose-normalization when dosenorm = TRUE. Accepts bare names or strings. Default is DOSE.

loq

Numeric value of the lower limit of quantification (LLOQ). BLQ imputation behavior on the DV / prediction layers is controlled by blq_mode and loq_method. Default is NULL.

loq_method

Method for handling BLQ data. See plot_dvtime() for option details. Default is 0 (no imputation).

blq_mode

One of "obs" (default) or "all". Controls which layers receive BLQ imputation: "obs" imputes the observed DV layer only, leaving PRED / IPRED untouched (mirrors plot_dvtime()); "all" additionally imputes the prediction layers, useful when the GOF visual should mirror an estimation engine that censored predictions to LLOQ. Has no effect when loq_method = 0.

cent

Character string specifying the central tendency measure. See plot_dvtime() for option details. Default is "mean".

dosenorm

Logical indicating if observed data points should be dose normalized. Default is FALSE. Requires variable specified in dose_var to be present in data.

ref

Numeric y-intercept for a horizontal reference line, or NULL for no reference line.

log_y

Logical indicator for log10 transformation of the y-axis. Also controls whether the caption reports arithmetic or geometric mean when show_caption = TRUE.

show_caption

Logical indicating if a caption should be shown describing the data plotted.

style

A ggstylekit::style_spec() controlling plot aesthetics. Defaults to style_gof(); view the defaults by running style_gof() with no arguments. Customize by passing style = style_gof(...), or restyle the returned plot with restyle_plot(). The DV/PRED/IPRED (and OBS) overlays are colored by the colors map keyed by those labels. The error bar cap width is the style's errorbar_width field; when unset it defaults to 2.5% of the maximum nominal time (ntime_var).

Value

A ggplot2 plot object

See also

Other goodness-of-fit: plot_gof_shown(), style_gof()

Examples

plot_gof(data_sad_pkfit, dv_var = ODV, dosenorm = TRUE)
#> Warning: Removed 169 rows containing non-finite outside the scale range
#> (`stat_summary()`).
#> Warning: Removed 169 rows containing non-finite outside the scale range
#> (`stat_summary()`).
#> Warning: Removed 169 rows containing missing values or values outside the scale range
#> (`geom_point()`).