Map - profile megastudy/zero imputation-related methods

[bobular]

See if we can find some speedup for the floating visualizations in the megastudy.

Particularly affects sample-based visualisations.

Here's one I made earlier...

Site: qa.vectorbase.org
Study: megastudy
Filters:

• Study.study_name LIKE 'Salt Lake' (e.g. filter with search box and select all)
• Collection start date: 2012 to 2022 (just select around the pink in the histogram, doesn't have to be exact)
  Donut marker overlay: Sample.species (default)
  Floater: Timeline, x=collection start date, y=Sample.unbiased specimen count (but other Sample variables with categorical proportions seem equally slow)

For me it was taking around 20 seconds to make a plot. I changed variable in the plot settings rather than change filters to trigger a reload, so that map markers, timeline data etc weren't being requested at the same time.

[d-callan]

we can make a separate ticket to profile plot.data, if needed. its been a while since thats been done.. but i think we should profile this first as its the more likely culprit and its never been done.