Output Venn diagram visualization for a set of selections

[Lambeaux]

Issue summary & value proposition

Visually represent the relationship between multiple selections so users know they are getting the data they expect.

Problem

Visualization is challenging because it's not always clear how to detect the difference between:

• Bad data
• User error
• An actual bug

When writing selections for filtering dependencies, seeing the relationship between multiple selections can be really useful so that follow-on analysis can be performed with more confidence.

Proposed solution

Provide another draw function that operates on selections or labeled selections:

;; Takes variadic args of filters and labels them using a stringified version of the filters
(draw-venn [:node "ddf/.*"] [:node ".*/.*catalog.*"] [:node ".*/.*spatial.*"])
;; Takes variadic args of alternating labels and filters
(draw-venn-labeled "ddf-nodes" [:node "ddf/.*"] "ddf-catalog" [:node ".*/.*catalog.*"] "everything-else" [:node ".*"])

Could also provide a counterpart that operates on features by aggregating the bundles into the appropriate buckets to assess how independent a set of features truly are:

(draw-venn-feature "feature-A" "feature-B" "feature-C")

[Lambeaux]

Helps us answer the hard data questions:

When someone doesn't get something, could the tail case just be that their graph doesn't have the requisite edges? Or the edges just aren't coming up in a search?

Hard to answer a question of omission without additional tools to peek into other aspects of the data.