"Modeling" is its own antonym
There are some words that are contronyms, with "sanction" being the most famous example (half the time meaning "allow" and the other half the time meaning "disallow"). "Modeling" is a contronym.
In the realm of data science, it usually means statistical modeling. This is entirely empirical. But as I wrote long ago on identifying causality, modeling can also mean constructing an engineering model that explains system behavior from first principles. The former is empirical and the latter is theory.
The difference is analogous to the two competing approaches to codifying grammar and language: descriptive vs. prescriptive. Without getting into a debate over grammar, when it comes to Data Science, the ultimate goal has always been prescriptive analytics, and actually identifying causality via modeling in the "theory/egnineering" sense of the word rather than the "empirical/statistical" sense of the word will lead to predictions less sensitive to overfitting and spurious correlations.
The use of "models" in Data Science is sometimes source of controversy, but it's important to distinguish which of the two senses is meant. Or it could be both senses, if statistics bear out a theoretical model.