Graphical Modeling and Causal Inference
Welcome
The Graphical Modelling and Causal Inference (GMCI) initiative provides an open repository of curated research resources for the statistical study of graphical models and causal inference. Its primary outputs are a moderated collection of datasets and a library of peer-reviewed, executable analysis notebooks, both archived through a dedicated Zenodo community with persistent digital object identifiers.
The initiative is part of the Mathematical Research Data Initiative (MaRDI), funded by the German National Research Data Infrastructure (NFDI), and is led by researchers from TU Munich, LMU Munich, and WIAS Berlin.
Scope and Purpose
Graphical models provide a principled framework for representing and reasoning about conditional independence structure and causal relationships among random variables. Causal inference addresses the identification and estimation of causal effects from observational and interventional data. Together, these fields underpin a growing body of methodology with applications across the natural sciences, medicine, economics, and the social sciences. Despite this breadth, the field has lacked a centralised, quality-controlled resource for sharing the datasets and reproducible analyses on which methodological progress depends.
GMCI was established to address this gap. The initiative moderates submissions of datasets and analysis notebooks from the wider research community, ensuring that each contribution meets documented standards for metadata completeness, reproducibility, and licensing transparency. Accepted submissions are indexed and made permanently accessible via Zenodo, enabling them to be cited in publications, integrated into benchmarking studies, and reused in downstream analyses.
Resources
The dataset collection contains curated datasets drawn from a range of application domains. Each entry is documented according to a standardised metadata schema that specifies, among other attributes, the data type, the causal or statistical task it supports, the ground truth availability, and the applicable licence. A supplementary overview of external data sources suitable for custom extraction is also provided, organised by domain.
The notebook library contains executable analyses that demonstrate methodology or present novel applications in graphical modelling and causal inference. Each notebook is developed from a common template, archived as a versioned Zenodo record with a citable DOI, and rendered for direct reading on this website. Contributions are accepted from authors across the academic community following a review process conducted by the GMCI moderators.
FAIR Principles
All resources in the GMCI collection adhere to the FAIR principles for scientific data management (Wilkinson et al. 2016). Resources are Findable through consistent use of persistent identifiers and standardised metadata. They are Accessible through open archiving on Zenodo under clearly specified licences. Interoperability is supported through the use of common data formats and documented schemas. Reusability is ensured through explicit provenance records and licensing information accompanying every submission. Researchers who use GMCI resources are asked to cite both the original creators and the community itself; citation details are provided on the About page.
Contributing
The initiative actively solicits submissions of datasets and analysis notebooks from the research community. Detailed instructions for preparing and submitting each type of contribution are provided on the Contributing page. The GMCI moderators review all submissions, may request revisions, and assist contributors throughout the process. Submission and archiving are free of charge.
Citation
When using resources from the GMCI community, please cite the original creators of the content and the community with the following reference.
@article{Mareis_Haug_Drton_2025,
title = {MaRDI's Zenodo Community for Graphical Modeling and Causal Inference},
volume = {2},
journal = {Proceedings of the Conference on Research Data Infrastructure},
author = {Mareis, Leopold and Haug, Stephan and Drton, Mathias},
year = {2025},
month = {Sep.}
}Upcoming Events
Members of the GMCI team will present at the following conferences.
| Event | Dates | Location |
|---|---|---|
| CLeaR 2026 | 03-06 April 2026 | Boston, United States of America |
| IMS Annual Meeting 2026 | 06-09 July 2026 | Salzburg, Austria |
| Announced soon | August 2026 |