In mvuorre/vmisc: Vuorre Miscellaneous

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

vmisc

vmisc is a small R package for convenience functions and plotting themes. You are probably here because you wanted to try one of the ggplot2 themes, in which case see below:

Installation

You can install vmisc from github with:

# install.packages("devtools")
devtools::install_github("mvuorre/vmisc")

ggplot2 themes

See below for examples of the themes in the vmisc package.

library(ggplot2)
library(gridExtra)
library(vmisc)
p <- ggplot(mtcars, aes(x=hp, y=mpg, col=as.factor(vs))) +
    geom_point() +
    geom_line() +
    facet_wrap("am", labeller = label_both, nrow=1)

theme_vmisc() is the base plot on which the other themes are built. It has three key options, illustrated below:

grid.arrange(
    p + theme_vmisc(),
    p + theme_vmisc(legend = "top"),
    p + theme_vmisc(grid = TRUE),
    p + theme_vmisc(facet_label = TRUE)
)

theme_vmisc(), with its three main options, replaces old functions theme_blog(), theme_poster(), and theme_pub(). Using these themes now gives a warning that they are deprecated:

p + theme_blog() + ggtitle("theme_blog()")

p + theme_poster(base_family = "Comic Sans MS") + ggtitle("theme_poster()")

p + theme_pub(base_size = 14) + ggtitle("theme_pub()")

theme_beamer() is tailored for Beamer pdf presentations, but accepts same three arguments as theme_vmisc():

p + theme_beamer() + ggtitle("theme_beamer()")

brms forest plot

brms_forest() is a convenience function for drawing forest plots from meta-analysis models fitted with brms. See ?brms_forest. The function is experimental and may contain bugs; if you find one, please leave an issue on GitHub.

Example use:

library(metafor)
library(dplyr)

d <- dat.bangertdrowns2004 %>%
    mutate(author = paste0(author, " (", year, ")"), sei = sqrt(vi)) %>%
    select(id, author, yi, sei) %>%
    slice(1:10)

To estimate a brms meta-analytic model on above data, run:

library(brms)

fit <- brm(yi | se(sei) ~ 1 + (1|id),
           data = d,
           cores = 4, control=list(adapt_delta = .99))
save(fit, file="tmp.rda")

load(here::here("tmp.rda"))

To draw the forest plot:

x <- brms_forest(data = d,
                 model = fit,
                 study = "id",
                 label = "author",
                 xlim = c(-2,2),
                 level = .99,
                 show_data = TRUE,
                 sort_estimates = F)

x + ylab("Effect Size") + scale_y_continuous(breaks = c(-1, 2))

New brmsfit method:

fit <- brm(yi | se(sei) ~ 1 + (1|author),
           data = d,
           cores = 4, 
           control=list(adapt_delta = .99))
save(fit, file="tmp.rda")