(13-05-2019, 01:04 AM)venam Wrote: I like how data science fits well with Unix, the idea of constructing and manipulating data sets.

it does, especially make comes in handy to both manage and document workflows.

there are some areas though where the "bare" unix is not sufficient, especially job control (batch systems required -- slurm is cool, makeflow is cooler) and reproducible results (hence scientists often using docker & singularity).