Decomposing galaxies with bang: an automated morphokinematic decomposition of the SDSS-DR17 MaNGA survey

From a purely photometric perspective galaxies are generally decomposed into a bulge+disc system, with bulges being dispersion-dominated and discs rotationally supported. However, recent observations have demonstrated that such a framework oversimplifies complexity, especially if one considers galaxy kinematics. To address this issue we introduced with the GPU-based code bang a novel approach that employs analytical potential-density pairs as galactic components, allowing for a computationally fast, still reliable fit of the morphological and kinematic properties of galaxies. Here we apply bang to the SDSS-MaNGA survey, estimating key parameters such as mass, radial extensions, and dynamics, for both bulges and discs of +10 000 objects. We test our methodology against a smaller subsample of galaxies independently analysed with an orbit-based algorithm, finding agreement in the recovered total stellar mass. We also manage to reproduce well-established scaling relations, demonstrating how proper dynamical modelling can result in tighter correlations and provide corrections to standard approaches. Finally, we propose a more general way of decomposing galaxies into 'hot' and 'cold' components, showing a correlation with orbit-based approaches and visually determined morphological type. Unexpected tails in the 'hot-to-total' mass-ratio distribution are present for galaxies of all morphologies, possibly due to visual morphology misclassifications.