as noted above, i think i know what the profile count is supposed to represent – we’re not in disagreement – but just to clarify, i’m not 100% sure of the exact series of steps that occurs in the black box to create that particular statistic. (i could look at the code to try to figure that out, i suppose, but i suspect it’s not worth the effort, and it’s probably going to open up more cans of worms.)
that said, i looked at one more possible source of the profile identification count, and i think i see where the main discrepancy between profile count and Android app count may be coming from. before, i made the assumption that counting active identifications would be equivalent to counting observations with active identifications, but that does not appear to be true.
if you count identifications (as noted in my earlier post), then this should tie to the Android app count: https://jumear.github.io/stirfry/iNatAPIv1_identifications.html?own_observation=false&user_id=mmmmbugs&verifiable=any.
however, if you count observations, then it looks like you’re more likely to get back to the total identification count noted on the profile page (except for maybe in some of the odd cases like what i noted in my previous post): https://jumear.github.io/stirfry/iNatAPIv1_observations?ident_user_id=mmmmbugs¬_user_id=mmmmbugs.
i can’t explain why counting active identifications is not equivalent to counting observations with active identifications, and i’m not going to dig into it right now (or maybe ever), but i suspect it’s mainly related to cases where identifications have somehow been orphaned, such as described here: https://forum.inaturalist.org/t/orphaned-identification-record-returned-from-api/30372. (if my suspicion ends up being true, then any difference would probably be related to observations that have been deleted since being identified.)
… that said, in cases where there’s a discrepancy, i still would tend to use the count on the Android app vs the count on the profile page, since the former is actually counting identifications (even if maybe some of them are orphaned at this point).
… in the end, i don’t think there’s a bug here exactly either. i might change the way the identification count on the profile page gets created just to make it consistent throughout the system, but it doesn’t seem like a particularly important change to make.