Launch Meta-Mar v4.0.2 • Features • Getting Started • Examples • Citation
Meta-Mar is an open-source statistical platform for conducting meta-analyses, developed for both research and educational purposes. The platform integrates AI assistance to guide users through methodological decisions and interpretation of results.
The tool supports a wide range of meta-analytic procedures, from basic effect size calculations to advanced heterogeneity assessment and publication bias evaluation. It is designed for researchers, educators, and students across various fields where evidence synthesis is valuable.
- Methodological Flexibility: Support for various outcome types (continuous, binary, correlations, effect size) and multiple model types
- Visualization Tools: Publication-quality forest plots, funnel plots, and other meta-analytic visualizations
- Heterogeneity Assessment: I², τ², and Cochran's Q statistics with subgroup analysis options
- Publication Bias Analysis: Egger's test, trim-and-fill analysis, and fail-safe N calculations
- Meta-Regression: Tools for exploring relationships between study characteristics and effect sizes
- AI Assistance: Interactive chatbot for methodological guidance and AI-powered report generation
- Built with R Shiny for an interactive web interface
- Leverages established R meta-analysis packages (meta, metafor, pimeta)
- Integrates AI capabilities using OpenAI's GPT models
- Full support for data import/export and result sharing
The easiest way to use Meta-Mar is through our hosted version:
To run Meta-Mar locally:
- Clone this repository
git clone https://github.com/yourusername/meta-mar.git
cd meta-mar- Install required R packages
install.packages(c("shiny", "meta", "pimeta", "readxl", "ggplot2", "dplyr",
"metafor", "DT", "httr", "markdown", "jsonlite",
"promises", "future", "shinyjs"))- Run the application
shiny::runApp()Here's a simple example of how to perform a meta-analysis using Meta-Mar's underlying code:
# Load libraries
library(meta)
library(metafor)
# Example data for continuous outcomes
data <- data.frame(
studlab = paste("Study", 1:5),
n.e = c(50, 45, 60, 90, 70),
mean.e = c(12.5, 13.2, 14.1, 15.0, 13.7),
sd.e = c(2.5, 2.8, 3.1, 2.9, 3.0),
n.c = c(48, 47, 55, 88, 65),
mean.c = c(10.2, 10.8, 11.5, 11.9, 10.5),
sd.c = c(2.4, 2.5, 3.0, 3.1, 2.8)
)
# Run meta-analysis
ma <- metacont(
n.e = data$n.e,
mean.e = data$mean.e,
sd.e = data$sd.e,
n.c = data$n.c,
mean.c = data$mean.c,
sd.c = data$sd.c,
studlab = data$studlab,
sm = "SMD",
method.smd = "Hedges",
method.tau = "REML",
method.random.ci = "HK",
prediction = TRUE
)
# Print results
summary(ma)
# Create forest plot
forest(ma,
common = FALSE,
random = TRUE,
prediction = TRUE,
lab.e = "Intervention",
lab.c = "Control",
comb.random = TRUE,
text.random = "Random effects model",
text.predict = "Prediction interval")For more complex examples and demonstrations, please visit our documentation.
meta-mar/
├── app.R # Main application file
├── global.R # Global settings and functions
├── user_summary.R # User summary functionality
├── documentation_MetaMar.R # Documentation content
├── assets/ # Images and static resources
├── example_data/ # Example datasets
└── README.md # This readme file
If you use Meta-Mar in your research, please cite it as follows:
Beheshti, A., Chavanon, M. L., & Christiansen, H. (2020). Emotion dysregulation in adults with attention deficit hyperactivity disorder: a meta-analysis. BMC psychiatry, 20, 1-11.
https://www.meta-mar.com
This project is licensed under the MIT License - see the LICENSE file for details.
Created by - Atriom Circle, Applied Intelligence Practice - For questions or support, please contact: a.beheshti@posteo.de