Structure and fragment-based lead-design (SBLD and FBLD) offer an efficient and rational route towards developing potent and selective small molecules. Facilities enabling crystallographic high-throughput fragment-based screening, such as XChem, have greatly increased the output of structural information available for SBLD/FBLD efforts. This influx of data has generated a need for computational tools to drive decision making for progressing fragment-hits cheaply and efficiently. My work focusses on three areas that address this need. 1) Generation of visualisation tools to enable human-driven analysis of structural data. 2) Semi-automated methods for  exploring chemical space and prioritising future experimental work. 3) Combined application of energetics, experimental and Deep Learning methods to understand protein-ligand complexes.