Which location-level factors drive risk — and how confidently can we say so?

The rigor flagship: one question answered as science, with an interpretable baseline standing next to the learned model.

Question

[State the precise question. What outcome, what unit of analysis, what counts as a "factor."]

What I expected

[Prior hypothesis before modeling — this is what makes it read as investigation, not a demo.]

Method & why this method

[The model used AND the justification vs. interpretable alternatives — regularized regression, GAM, tree ensemble + SHAP. Pre-empt "why a black box?"]

Baselines

[Interpretable baseline result side-by-side with the learned model. This comparison is the credibility.]

Validation

[How factor-importance estimates were validated; stability across resamples / regions.]

Factor-importance map

[Visualization slot — importance by location. The payoff visual.]

What surprised me

[A genuine surprise signals real engagement with the data.]

Limitations — what this can't tell you

[Confounding, ecological fallacy, data gaps. Stating limits is investigator-grade.]

What's next

[The next question this opens.]