Developing renewable energy such as tidal turbines requires in-depth assessment of a potential project site to understand suitability, potential energy production as well as impact on the environment. The exploitability of a site is mainly assessed by a combination of extensive field surveys with numerical modelling, which is expensive. Due to budget limitations, critical financial and technical decisions are made on a restricted sample of data leading to high level of risk and uncertainties. Here we aim to mitigate the issue of data scarcity by fusing established tidal flow analysis techniques with machine learning tools. The new tool will 'learn', from verified gauge data, the best way to temporally extend short-duration spatial survey data to make maps of tidal potential that can directly inform either more spatially targeted surveying, or assessments for optimal siting of tidal stream devices. The tool aims to make surveying potential sites cheaper by targeted adaption of the survey campaign and more robust analysis of the data than is currently practiced. This is a proof-of-concept study. The outcomes include assessing whether the tool has sufficient commercial merit to be developed further via a NERC follow-on call.