The main goal of the work presented in this thesis is to develop and apply computational tools for NMR crystallography, namely the combination of theoretical and computational methods with experimental solid-state NMR and powder X-ray diffraction. The Ab Initio Random Structure Searching (AIRSS) approach was applied to predict the structure of an organic solid, maminobenzoic acid (mABA). Assessment of candidate structures was carried out against experimental powder X-ray diffraction data and solid-state NMR data. A successful Le Bail refinement and a good agreement between calculated and experimental NMR chemical shifts was achieved for some of the lowest energy candidate structures, showing that two polymorphs, forms-III and IV, had been identifed. In further work a computational framework to identify how the local environment affects the NMR shieldings has been introduced. It is demonstrated that the majority of chemical shift variation for protons comes from long-range current contributions whereas for heavier atoms it is the short-range term that dominates. Magnetic Shielding Contributions Field (MSCF) plots are introduced as a visualisation tool for identifying the contributions to the magnetic shieldings from ring-currents and hydrogen-bonds. Next, an experimental and theoretical analysis of the polymorphic active pharmaceutical ingredient (API) furosemide is presented, employing the computational tools developed to identify local environment effects. Lastly, we highlight computational tools developed to process and visualise the output of ab initio calculations of NMR parameters.