Maximally informed Bayesian modelling of disc galaxies

Dissecting the underlying structure of galaxies is of main importance in the framework of galaxy formation and evolution theories. While a classical bulge+disc decomposition of disc galaxies is usually taken as granted, this is only rarely solidly founded upon the full exploitation of the richness of data arising from spectroscopic studies with integral field units. In this work we describe a fully Bayesian estimation method of the global structure of disc galaxies which makes use of the wealth of photometric, kinematic, and mass-to-light ratio data, and that can be seen as a first step towards a machine-learning approach, certainly needed when dealing with larger samples of galaxies. Ours is a novel, hybrid line of action in tackling the problem of galactic parameter estimation, neither purely photometric nor orbit-based. Being rooted on a nested sampler, our code, which is available publicly as an online repository, allows for a statistical assessment of the need for multiple components in the dissecting process. As a first case-study the GPU-optimized code is applied to the S0 galaxy NGC-7683, finding that in this galaxy a pseudo-bulge, possibly the remnant of a bar-like structure, does exist in the center of the system. These results are then tested against the publicly available, orbit-based code DYNAMITE, finding substantial agreement.