Consider storing data in a separate repository

[MarionBWeinzierl]

As noted in #185, it would be worth to think about storing datasets for tests, profiling etc. in a separate data repository like figshare or Zenodo, rather than in this code repository.

When doing this, a mechanism should be implemented to easily access this data, i.e., a script to download it, place it in the right place, and potentially do preprocessing steps.