Consider tracking trends of data metrics over time

[mikix]

"How have our quality scores improved over time?" is an interesting question to be able to ask.

In Cumulus-land, we might be able to get by with an aggregator-side feature that tracks uploads and tags them with the time.

In local-data-land, maybe we can do something to help with examining multiple runs of the data. Separate buckets for run data and then a unified view of that? Tagging the tables with run time?

How much of this lives in the data-metrics study vs a visualizer repo?