Since the fix, I haven’t really dug into exactly which observations are shown or hidden on the taxon maps, probably because there are many fewer inconsistencies than before.
I agree with @bouteloua that I do not want to see spam, copyright infringement or opted out maverick IDs. I do think there are likely other observations that are casual for reasons other than the lack of media. @pisum raises the issue of an observation that is marked “as good as it can be” at family level. That seems like something I would want to see on the Family level taxon map, and likely to be a real issue with taxa such as invertebrates, lichens, etc.