itbound Demo

Paper | GitHub | PyPI

Upload a CSV to compute data-driven lower and upper causal bounds under unmeasured confounding, or leave the upload blank to run the canonical IHDP example. For a toy uploaded CSV, a common schema is treatment=a, outcome=y, and covariates=x1,x2.

CSV upload (optional; blank = canonical IHDP example)

Treatment column

Outcome column

Covariate columns (comma-separated)

Ribbon x-axis column

Recommended: propensity for the paper-aligned IHDP view. You can also try a raw covariate such as x4.

Aggregation mode

paper_adaptive_k fixed_k_endpoint tight_kth

Write HTML report

Bounds preview

Interval width histogram

Binned ribbon plot

Download artifact bundle