If you have some data, you can prove that it is consistent with a model or that your methodology will extract useful information from noisy data assuming it fit

If you have some data, you can prove that it is consistent with a model or that your methodology will extract useful information from noisy data assuming it fits your model of it. Especially in physics and number-crunching tasks like these, there is enough overlap with mathematics to have plenty of opportunity for proving steps of your process correct. Also mathematics does not have a monopoly on employing pure reason.