A data processing toolbox for agile scientific research
Thomas Edwards | Saturday 10:15 | Room B
An in-house data processing API/toolbox has been developed for research scientists at a nuclear fusion startup. The aim of the toolbox is to help scientists across the company increase the efficiency, reproducibility and robustness of processing data from physical experiments and numerical simulations. The toolbox is designed to provide a unified workflow for both simulations and experiments.
To achieve this goal, where user python experience ranges from beginner to professional, we undertook an iterative architectural design and development strategy, with a high level of user engagement. In this talk, we outline our approach for gathering requirements, a selection of user stories, and the architectural decisions that enabled the toolbox to achieve its goals with a high user uptake.