About Open Causal

Let's make FAIR assumptions!

Causal graphs are essential to causal inquiry in many fields, including epidemiology, economics and computer science. Using the graph to present causal assumptions is strongly recommended, yet their adoption remains limited. Even when reported, they are only available in image format. Currently, there is no platform for researchers to share causal graphs, which hinders their reusability and makes it difficult to build upon existing knowledge.

We introduce Open Causal, a platform to facilitate sharing, discussion, and exploration of causal graphs. Open Causal will function as an open registry for causal graphs, where researchers can publish their graphs in machine-readable format with an open licence and receive a DOI. Other researchers can search, comment, clone and modify existing graphs, fostering an open discussion on assumptions and enabling reuse of the models. The platform facilitates community-driven curation and iterative evolution of graphs. An API allows integration with popular statistical packages in Python and R. Additionally, the platform supports empirical validation, enabling researchers to test causal assumptions against their datasets without uploading sensitive data. The platform also integrates AI-powered tools for advanced use cases.

Open Causal transforms once immutable supplementary materials into interactive open research objects. It fosters the adoption of FAIR principles (findability, accessibility, interoperability, reusability) in causal inference.