gg_coefsplot: Caterpillar plots of regression coefficients from a boral...
In mbedward/ggboral: View BORAL model results with ggplot

Description Usage Arguments Examples

View source: R/coefsplot.R

Constructs horizontal line plot showing 95 intervals, otherwise known as a caterpillar plot, for the column-specific regression coefficients corresponding to the covariates in X fitted in the boral model. Median coefficient values are shown as a point on each interval.

gg_coefsplot(
  model,
  X.labels = NULL,
  linesize = 1.5,
  pointsize = 3,
  palette = "Greys",
  single.colour = "black"
)

`model`	A boral model fitted with one or more latent variables.
`X.labels`	An optional vector of text strings to use as labels for the X variables (graph facet labels). If a named vector is provided (safest), the names will be used to match labels to the X variable names in the fitted model. If `NULL` (default), X variable names are taken from the model.
`linesize`	Thickness of the line segments drawn for intervals (passed to ggplot as the `size` argument for `geom_segment`).
`pointsize`	Size of points representing median values (passed to ggplot as the `size` argument for `geom_point`). Set to zero if you do not want to display points.
`palette`	The name of the colour palette to use when colouring intervals to indicating where intervals lie in relation to zero. This should be one of the sequential palettes recognized by `scale_colour_distiller`. The default is "Greys". If set to `NULL` or `""` (empty string), all intervals will be drawn in the same colour, as specified by `single.colour`.
`single.colour`	The name or hex code of a single colour to use for all intervals. This argument only applies if `palette` is set to `NULL` or `""`. The default single colour is black.

# This example is based on that for the coefsplot function
# in the boral package and requires an example dataset from the
# mvabund package.
library(boral)
library(ggboral)

data(spider, package = "mvabund")
y <- spider$abun
X <- scale( spider$x )

# Warning - these settings are only to make the example run quickly.
# Don't use them for a real analysis!
example.control <- list(n.burnin = 10, n.iteration = 100, n.thin = 1)

spiderfit_nb <- boral(y, X = X,
                      family = "negative.binomial",
                      lv.control = list(num.lv = 2),
                      mcmc.control = example.control)

gg_coefsplot(spiderfit_nb)

# Display plot again with a white background
last_plot() + theme_bw()

# Re-do plot using a different palette for the intervals
# (see help page for ggplot2::scale_colour_distiller for the
# available palette names)
#
gg_coefsplot(spiderfit_nb, palette = "Blues") +
  theme_bw()

# Draw all intervals with a single colour
# (the palette argument must be set to NULL otherwise
# it overrides the single.colour argument)
#
gg_coefsplot(spiderfit_nb, palette = NULL, single.colour = "darkred") +
  theme_bw()