This package is a collection of experimental functions to scale meta-analyses. The package is still under development, the functions are not yet fully tested and the results should be interpreted with caution. Some functions may be unstable if several levels of groups are present in the dataset.
To install the package, run the following:
devtools::install_github("github.com/thomazfabrin/metaPipeline")
The main advantage of metaPipeline is the ability to create a fully customizable forest plot. The package also includes functions to clean data, run a random-effects model, and create a summary table.
As an example of how to use the package, run the following:
library(metafor)
group <- "outcome"
group_plot <- "group1"
order <- c("OS", "DFS")
coordinates <- metaPipeline::clean_data(dat.riley2003) |>
metaPipeline::auto_rma_uni(group = group) |>
metaPipeline::rma_list_to_summary_df() |>
metaPipeline::mining_diamonds(
main_group = group, order = order,
diamond_height = 0.25)
plot <- metaPipeline::plot_diamonds(
coordinates_df = coordinates,
title = "dat.riley2003 from metafor package",
show_grids_v = TRUE, xlab = "SMD", group1 = group_plot,
order = order, vline_ref = 0, adj1 = 1.2, adj2 = 16,
ci_interval = TRUE) +
ggplot2::theme(legend.position = "none")
ggplot2::ggsave(
filename = "figures/plot.png",
plot = plot, width = 8, height = 2,
units = "in", dpi = 300)
You should get a plot similar to the one below:
If you detect any bugs or errors, or have any suggestions, please open an issue. I hope this package can be useful to you.
TF