Now I know that https://datproject.org/ has solutions for exactly this problem and more, with a focus on usability and the big-lump-o-data scientific dataset.