Structure determination of molecular solids through NMR crystallography relies on the generation of a comprehensive set of candidate crystal structures and on the comparison of chemical shifts computed for those candidates with experimental values. Exploring the polymorph landscape of molecular solids requires extensive computational power, which leads to a significant bottleneck in the generation of the set of candidate crystals by crystal structure prediction (CSP) protocols. Here, we use a database of crystal structures with associated chemical shifts to construct three-dimensional interaction maps in molecular crystals directly derived from a molecular structure and its associated set of experimentally measured chemical shifts. We show how the maps obtained can be used to identify structural constraints for accelerating CSP protocols and to evaluate the likelihood of candidate crystal structures without requiring DFT-level chemical shift computations.