Observed DataCausal Inference for real-world applicationsCausal Relationships

Get started with causal inference

causy is a package for discovering and quantifying causal relationships in observed data. Explore causality using our guided processes to understand the assumptions of different causal inference methods and visualize the results with causy UI. Use our pre-configured pipelines to run causal discovery algorithms and causal effect estimation methods.

Explore your results

causy workspaces allows you to compare experiments with different methods and hyperparameters. This allows you to test the robustness of your results.

causy UI allows you to visualize your experiments and gain further insights in the causal discovery process. You can then share your results with your coworkers.

Use causy in your software

By its modular architecture, causy enables easy modification of pipeline steps, adjustment of algorithms, and reuse of individually configured pipelines in your code. Built with PyTorch, causy leverages GPU acceleration and supports serialization.