examples

StatsPAI Examples

These examples are short, offline scripts for reviewers and new users. They use the teaching datasets bundled with statspai, so they do not download data or require network access after installation.

From a source checkout:

python -m pip install -e ".[dev,plotting]"
python examples/card_iv.py
python examples/did_mpdta.py
python examples/rd_lee.py
python examples/synth_prop99.py
python examples/gmethods_timevarying.py

Or after installing the released package:

python -m pip install statspai
python examples/card_iv.py

The scripts cover canonical causal-inference designs:

• card_iv.py - instrumental variables using Card (1995).
• did_mpdta.py - staggered difference-in-differences using mpdta.
• rd_lee.py - sharp regression discontinuity using Lee (2008).
• synth_prop99.py - synthetic control using California Proposition 99.
• gmethods_timevarying.py - g-methods (parametric g-formula + marginal structural model) for time-varying confounding, the signature problem of modern causal epidemiology. Uses a self-contained simulation, so it needs no bundled dataset.