Some thoughts on Statistical Modeling

I'd like address some general questions about statistical modeling: What is it?, Why do we do it? How do we do it?

Much of statistical modeling, including logistic regression fits in a single framework.

Three (non-exclusive) reasons we might want to model.

The precise methods we use depends on context, but there are some necessary ingredients.