Meta-Mar

A free, web-based meta-analysis platform built with R Shiny, integrating comprehensive statistical methods with AI-powered methodological guidance.

Live Application: https://www.meta-mar.com

Summary

Meta-Mar enables researchers to conduct meta-analyses without programming knowledge or software installation. The platform provides:

• Support for continuous, binary, correlation, and pre-calculated effect size data
• Fixed-effect and random-effects models with 8 heterogeneity estimators
• Publication bias assessment (funnel plots, Egger's test, trim-and-fill, fail-safe N)
• Subgroup analysis and meta-regression
• AI-powered interpretation guidance via GPT-4 integration
• Publication-quality visualizations

Since 2020, Meta-Mar has been used by over 5,800 researchers across 120+ countries and cited in 200+ peer-reviewed publications.

Installation

Requirements

• R >= 4.0.0
• RStudio (recommended for local development)

Dependencies

install.packages(c(
  "shiny",
  "metafor",
  "meta",
  "ggplot2",
  "dplyr",
  "DT",
  "shinythemes",
  "shinyjs",
  "httr",
  "jsonlite"
))

Running Locally

git clone https://github.com/mirzafarangi/meta-mar.git
cd meta-mar

shiny::runApp()

The application will open in your default browser at http://127.0.0.1:xxxx.

Usage

For complete documentation, see: https://www.meta-mar.com/documentation

Data Input

Meta-Mar accepts CSV files or manual data entry for:

Data Type	Required Columns
Continuous	study, n.e, mean.e, sd.e, n.c, mean.c, sd.c
Binary	study, event.e, n.e, event.c, n.c
Correlation	study, cor, n
Pre-calculated	study, TE, seTE

Example

# Example continuous outcome data structure
data <- data.frame(
  study = c("Study A", "Study B", "Study C"),
  n.e = c(50, 75, 60),
  mean.e = c(12.5, 13.2, 11.8),
  sd.e = c(3.2, 2.9, 3.5),
  n.c = c(48, 72, 58),
  mean.c = c(10.1, 10.8, 9.9),
  sd.c = c(3.0, 3.1, 3.3)
)

Statistical Methods

Effect Size Measures

• Continuous: Standardized Mean Difference (SMD), Mean Difference (MD), Ratio of Means
• Binary: Odds Ratio (OR), Risk Ratio (RR), Risk Difference (RD)
• Correlation: Fisher's z transformation

Heterogeneity Estimators

REML, DerSimonian-Laird, Paule-Mandel, Maximum Likelihood, Empirical Bayes, Sidik-Jonkman, Hedges, Hunter-Schmidt

Confidence Interval Methods

Classic (Wald), Hartung-Knapp-Sidik-Jonkman (HKSJ), Kenward-Roger

Publication Bias

• Funnel plot visualization
• Egger's regression test
• Begg's rank correlation test
• Trim-and-fill adjustment
• Fail-safe N (Rosenthal, Orwin, Rosenberg)

Project Structure

meta-mar/
├── app.R              # Shiny UI and server logic
├── global.R           # Global functions, settings, demo data
├── user_summary.R     # Session management utilities
├── data/
│   └── statistics.json
├── paper.md           # JOSS paper
├── paper.bib          # References
├── CONTRIBUTING.md    # Contribution guidelines
├── LICENSE            # MIT License
└── README.md

Core Dependencies

Meta-Mar builds on established R packages for meta-analysis:

• metafor (Viechtbauer, 2010): Core meta-analytic computations
• meta (Balduzzi et al., 2019): Additional methods and visualizations
• shiny (Chang et al., 2024): Web application framework

Validation

Computational accuracy has been validated against Cochrane RevMan 5.4. See the accompanying paper for detailed comparison results.

Privacy

• No user registration required
• Uploaded data stored in session memory only
• Automatic deletion when session ends
• No persistent storage of user data
• AI interactions not permanently stored

Full privacy policy: https://www.meta-mar.com/privacy

Citation

If you use Meta-Mar in your research, please cite:

@software{beheshti2026metamar,
  author = {Beheshti, Ashkan and Sazmand, Hassan and Chavanon, Mira-Lynn and Christiansen, Hanna},
  title = {Meta-Mar: An AI-Integrated Web Platform for Meta-Analysis},
  year = {2026},
  url = {https://github.com/mirzafarangi/meta-mar},
  version = {4.0.2}
}

Contributing

Contributions welcome. Please:

1. Fork the repository
2. Create a feature branch
3. Follow R tidyverse style guide
4. Test with known datasets
5. Submit pull request

License

MIT License. See LICENSE for details.

Contact

• Website: https://www.meta-mar.com
• Issues: GitHub Issues
• Corresponding Author: Ashkan Beheshti (a.beheshti@posteo.de)

Acknowledgments

Development supported by Philipps-Universität Marburg, Department of Psychology.