Classification stats - evidence key filter and stats

Not infrequently, I want to see how people have actually filled in evidence keys.

For instance: how have people filled out the "splicing_assertion"? How many times in say, a P/LP?

I think it would be useful to be able to somehow filter what classifications are returned (say, via the standard grid + filters)

Then, select what evidence keys you are interested in. Either via a multi-select auto complete or maybe checkboxes (I generated one dynamically via get_mult_choice_form)

Then display value counts, or maybe even things like mean/graphs of distribution for fields that are int/floats etc

Start with the standard grid and filters entry point, then inspect how get_mult_choice_form is used for dynamically generated choices. Define the scope for filtering classifications by evidence key and presenting value counts or numeric distributions; done should include an agreed interaction and representative results for the requested evidence fields.

Written by the indexing model from the issue text.