Data provenance or processing breadcrumbs

Nobody has claimed this yet.

**Feature request**
It would be cool if we had a way to easily and confidently state, "this is the data that was used to create this plot".

This came up today during a sync meeting where @mtauraso was demo'ing visualization of a umap output. And the question arose, what data is being used to produce this plot? The config system supported being able to backtrack to the original data, but it was laborious.

@aritraghsh09 noted that he is having to do the same thing but in his case, it is for many different runs started in parallel. i.e. testing many different hyperparameters and producing outputs in many different result directories. It become laborious to backtrack for each to see what parameters or datasets were used to produce the final output. MLFlow helps with config changes, but it doesn't track the original input data.

**What would b cool to have**
Perhaps a simple visualization? Or some obvious way to tag output? Perhaps a breadcrumb system that shows the full processing history to get to the current results? Being able to "undo" or backtrack and then branch off with a new config would be interesting to support.

No files, tests, or entry points are named. Start by clarifying the provenance requirements and examining the existing config system and result directories; the issue does not yet define what behavior would count as done.

Written by the indexing model from the issue text.