Hubbard U and Hund J values provide a measure of the self-interaction between correlated electrons, and are crucial parameters in the formalism of density functional theory with a "plus U " correction, known as DFT+U and DFT+U +J. The linear response (LR) methodology has proven to be a computationally effective and self-contained method for computing accurate U and J values. This study provides a high-throughput computational analysis of the U and J values applied to transition metal d-electron states in a representative set of magnetic transition metal oxides (TMOs). In a less conventional pursuit, over two hundred system-specific U and J corrections are calculated for oxygen 2p on-site occupations. The distributions of values are analyzed for structures containing manganese, iron and nickel-containing compounds, with sample sizes of over 150 for each species. In addition, periodic tables of U and J values are presented for transition metal and oxygen species from a combined data set of over 800 TMO compounds. Our work provides a frame of reference for researchers who utilize DFT+U to study TMO materials, and to gain insight into the distribution of U values that may be relevant to their applications. An atomate workflow is presented for calculating U and J values automatically on massively parallel supercomputing architectures. To validate this method, the spin-canting magnetic structure and unit cell parameters of a known multiferroic, olivine LiNiPO4, are predicted using the computed Hubbard U and Hund J values for Ni-d and O-p on-site occupancies, and are compared with experiment. It was confirmed that in addition to Ni-d U values, applying a separate Hund J to Ni-d has a strong effect on Ni-moment canting angle. Additionally, including a O-p U value results in a significantly improved agreement between DFT-computed lattice parameters and experiment.