General framework
01Causal Inference Engine
Given a query, available data, and a causal model, the engine asks whether the target is identifiable—and, when it is, returns an estimand.
QQuery
DData
MModel
Framework: Bareinboim & Pearl (2016)
We believe
Science advances when researchers embrace causal methods.
Individuals strengthen communities through causal literacy.
Everyone can learn causal methods.
Two complementary approaches
What can be learned from data?
One establishes what is learnable in principle. The other turns that reasoning into a practical sequence of study decisions.
Applied workflow
02The Causal Roadmap
The Causal Navigator applies that logic to a real study, keeping the question, assumptions, target, analysis, and interpretation connected.
- 01Question
- 02Assumptions
- 03Estimand
- 04Analysis
- 05Interpretation
The ecosystem
Choose your way in.
Start from the work in front of you: design a study, build intuition, or investigate the evidence.
010203
Design
Causal Navigator
Build a transparent causal study, from the question through interpretation.
Open the NavigatorLearn
Methods in motion
Develop causal intuition through practical notes and worked examples.
Explore learningInvestigate
Familiar workbench
Connect the literature, evidence, and assumptions around a question.
Open the workbenchCurrent work · ACIC 2026
SHAP has short memory. Causal SHAP remembers the DAG.
Our ACIC 2026 work asks what changes when feature attribution respects the causal structure that generated the data—not merely the associations the model can see.
Read the project