Scanning electron diffraction offers researchers relatively easy access to the nanoscale and is often used to map sample properties in crystalline materials. Data quality can often be improved by introducing double conical rocking of the beam (commonly called precession). In this thesis, routes to improvement for two key areas within the field - orientation mapping and strain mapping - are demonstrated, allowing for significant increases in both throughput and accuracy. Comparisons are made between different electron-based techniques including the extremely popular electron backscatter diffraction method. We present results investigating the changes that occur under a nano-indent in the ultra-hard material cubic boron nitride, illustrating the well-suitedness of scanning electron diffraction to the investigation of such highly deformed materials. Our findings indicated that a lobe like orientation pattern is formed, similar to that seen by other researchers conducting experiments with copper, a substantially softer material. Following this we provide a number of technique agnostic machine learning approaches for working with orientations, a feature of crystalline samples that is often of interest to researchers as texture can have a profound effect on material properties. Several other samples, both studied and potential are then considered in the closing chapters, giving the reader a broad sense of the range of investigations that scanning electron diffraction can facilitate.