i think that chemistry and the biological sciences can get pretty "empirical" at times. like chemical reactions/processes which are done in a specific way simply because experience shows that that's the best way to do it. or pharmaceuticals whose efficacy is supported mostly by experience, and we have just vague ideas for mechanisms of action.
sure these cases are usually not 100% empirical, there is some theory... but contrast it with, say mechanisms in machines or electronic circuits where you can usually correctly predict the effect on the behavior of the variation of whatever parameter you want.
i wish i could come up with a lot of examples but i don't think they're hard to find. i'll just say the following: just saw a thread on HN about how exposure to certain pesticides may cause Parkinson's disease. there is some theory behind it (the molecule (paraquat) is similar to another known one (MPTP) which causes Parkinson's disease immediately upon ingestion), but that's it, the rest of the evidence is like empirical correlations between proximity to agricultural regions and higher prevalence of Parkinson's disease.
if you think that's not empirical enough, again, compare it with whatever electronic circuit where you can precisely predict the effect of changing most of the parameters of the circuit.
just my uneducated opinion...
EDIT: just remembered one thing. relevant passage from Ashcroft & Mermin "Solid-State Physics":
> Justifying the semiclassical model in detail is a formidable task, considerably more difficult than justifying the ordinary classical limit for free electrons. In this book we shall not offer a systematic derivation. Our emphasis instead will be on how the semiclassical model is used. We shall therefore simply describe the model, state the limitations on its validity, and extract some of its major physical consequences.
> The reader who is dissatisfied with the very incomplete and merely suggestive bases we shall offer for the semiclassical model is urged to examine the broad array of mysteries and anomalies of free electron theory that the model resolves. Perhaps a suitable attitude to take is this: If there were no underlying microscopic quantum theory of electrons in solids, one could still imagine a semiclassical mechanics (guessed by some late nineteenth-century Newton of crystalline spaces) that was brilliantly confirmed by its account of observed electronic behavior, just as classical mechanics was confirmed by its accounting for planetary motion, and only very much later given a more fundamental derivation as a limiting form of quantum mechanics.
so the theory of electrical conduction in metals or semiconductors, i'm not sure it has been fully justified from more fundamental physics, or to what extent it's "empirical". the same probably goes for chemical reactions, which in principle could be entirely explained on the basis of QM, but probably aren't.