Draw diagnostic parades — diagnostic

These are short-hand functions to quickly draw diagnostic parades.

lin_plot() plots the models' residuals against (by default) their corresponding fitted values. The residuals can also be plotted against a specific predictor if the predictor parameter is set. This function is useful for checking the linearity assumption.

var_plot() plots the models' absolute residuals against (by default) their corresponding fitted values if it is fed an object generated using parade(). The absolute residuals can also be plotted against a specific predictor if the predictor argument is set. If it is fed an object generated using parade_summary(), it plots the sample standard deviation of the residuals per cell. This function is useful for checking the constant-variance assumption.

norm_qq() and norm_hist() plot normal quantile-quantile plots and histograms of the models' residuals, respectively. This function is useful for checking the normality assumption.

Usage

lin_plot(parade, predictor = NULL, rank = FALSE)

var_plot(parade, predictor = NULL, rank = FALSE)

norm_qq(parade)

norm_hist(parade, bins = 30)

Arguments

parade: The name of an object generated using parade(). For var_plot(), objects generated using parade_summary() are also accepted.
predictor: The name of a variable in the parade object against which the residuals should be plotted. If this parameter isn't specified (default), the residuals will be plotted against their respective fitted values.
rank: Should the values along the x-axis be converted to ranks (TRUE) or not (FALSE, default)? When used, ties are broken randomly. This may be useful when the raw values are concentrated in certain regions along the x-axis, making it difficult to discern relevant patterns.
bins: How many bins should the histograms contain? Defaults to 30.

Examples

# A simple regression model
m <- lm(mpg ~ disp, data = mtcars)

# Generate parade and check linearity
my_parade <- parade(m)
lin_plot(my_parade)
#> `geom_smooth()` using method = 'loess' and formula = 'y ~ x'

reveal(my_parade)
#> The true data are in position 1.

# Regenerate parade and check constant variance
my_parade <- parade(m)
var_plot(my_parade)
#> `geom_smooth()` using method = 'loess' and formula = 'y ~ x'

reveal(my_parade)
#> The true data are in position 15.

# Regenerate parade and check normality
my_parade <- parade(m)
norm_qq(my_parade)

norm_hist(my_parade)

norm_hist(my_parade, bins = 10)

reveal(my_parade)
#> The true data are in position 12.

# Example with gam
library(mgcv)
#> Loading required package: nlme
#> 
#> Attaching package: ‘nlme’
#> The following object is masked from ‘package:dplyr’:
#> 
#>     collapse
#> This is mgcv 1.9-4. For overview type '?mgcv'.
m.gam <- gam(mpg ~ s(disp) + wt + s(qsec, by = am), data = mtcars)
my_parade <- parade(m.gam)
lin_plot(my_parade)
#> `geom_smooth()` using method = 'loess' and formula = 'y ~ x'

lin_plot(my_parade, predictor = "wt")
#> `geom_smooth()` using method = 'loess' and formula = 'y ~ x'

lin_plot(my_parade, predictor = "qsec")
#> `geom_smooth()` using method = 'loess' and formula = 'y ~ x'