plotForest: Forest plot

In lwjohnst86/rstatsToolkit: Bundles up all my most used functions for doing statistical analysis

Description

Generate a forest plot without the traditional side table.

Usage

plotForest(data, coefficient = "estimate", y.axis.variables = "indep",
  confid.interval = c("conf.low", "conf.high"), pvalue.factor = NULL,
  groups = NULL, y.axis.label = "Exposures",
  x.axis.label = "Beta estimates")

Arguments

data	Dataset for the forest plot.
coefficient	The column that contains the beta estimate/coefficient.
y.axis.variables	The column with the exposure variables that will be placed on the y-axis of the forest plot.
confid.interval	A vector that contains the lower and upper confidence interval.
pvalue.factor	The column that contains the p-value in the form of a factor variable (ie. with levels such as '>0.05' and '<0.05').
groups	The variable to split the plot up, as a formula (var1 ~ var2, or ~ var2, etc).
y.axis.label	The y-axis label.
x.axis.label	The x-axis label.

Details

Create a forest plot, with a dot and confidence line, though without the usual side table that contains the raw data values. If the pvalue.factor argument is supplied, the dots and confidence lines increase in size and opacity as significance increases. If groups is also supplied, the forest plot will be split up for each grouping. Thus, a large amount of information on the results can be provided in a fairly small amount of space.

Value

A forest plot

Author(s)

Luke W. Johnston

Examples

## Not run: 
data(state)
ds <- data.frame(state.region, state.x77)
geefit <- loopGEE(ds, c('Income', 'Frost'), c('Population', 'Murder'), 'state.region')
  filter(term == 'independent') %>%
filtered <- dplyr::filter(geefit, term == 'independent')

plotForest(filtered)
plotForest(filtered, groups = ' ~ dep')
plotForest(filtered, pvalue.factor = 'f.pvalue', groups = ' ~ dep')

## End(Not run)

lwjohnst86/rstatsToolkit documentation built on May 21, 2019, 9:15 a.m.