Reticular chemistry provides materials designers with a practically infinite playground on different length scales. However, the space of all plausible materials for a given application is so large that it cannot be explored using a brute-force approach. One promising approach to guide the design and discovery of materials is machine learning, which typically involves learning a mapping of structures onto properties from data. To advance the data-driven materials discovery of metal-organic frameworks (MOFs) for gas storage and separation applications we provide a dataset of diverse gas separation properties (CO2, CH4, H2, N2, O2 isotherms); H2S, H2O, Kr, Xe Henry coefficients (computed using grand canonical Monte-Carlo with classical force fields) as well as parasitic energy for carbon capture from natural gas and a coal-fired power plant (computed using a simple process model) for the relaxed structures in the QMOF dataset with their DDEC charges.