Automated Statistical Analysis

Tea is a high-level language and runtime system
for automatically selecting and executing valid statistical tests.

Tea aims to lower the barrier for conducting valid, reproducible statistical analyses for novices and experts alike. Tea achieves this by eliciting users' domain knowledge and hypotheses and verifying properties of the data to infer applicable, valid statistical tests.


Peruse through community examples.

Want to be featured? We are always looking to feature people who have used Tea. Please get in touch if you'd like your work featured!

Research publications and presentations

Read our paper on Tea that was published at ACM UIST 2019. Prefer audio or video instead? Check out the video recording of the presentation. October, 2019.

Watch the longer research talk on Tea at Microsoft Research. August, 2019.


The development of Tea is led by Eunice Jun (UW), Rene Just (UW), Jeffrey Heer (UW), and Emery Berger (UMass Amherst).

Tea benefits from significant contributions from its active community, including Marcin Malinowski, Melissa Birchfield, and others.

Previous contributors and collaborators include Maureen Daum (UW), Jared Roesch (UW), Sarah Chasins (UC Berkeley), and Katharina Reinecke (UW).

If you would like to collaborate, please get in touch!

Join us in the Tea Room to discuss issues, plan future development, share your experiences, and to just chat!