1. usually ~/.config/PROGRAM_NAME/

You probably want to look into $XDG_CONFIG_HOME and the freedesktop spec documentation if you are interested in more of where these standards come from:
https://specifications.freedesktop.org/basedir-spec/latest/ar01s02.html

(and more generally)

https://www.freedesktop.org/wiki/Specifications/

obviously good for the bit, but the serious reason why this exists is likely because it's a modern gtk app -> it's meant for gnome users / made by a gnome users -> gnome is all in on freedesktop/flathub -> these users can all expect to have flatpak availability as a common method of distribution -> therefore, release it as a flatpak, so these users who already have flatpak can more easily manage their appimage-only programs

Worth noting that some of this isn't always 💯 possible. If some instances have blocked other instances where users have interacted with a post, your numbers and their numbers will diverge. The host of the content generally will have the most complete numbers, but it might be possible for that to be wrong if the server doesn't have authorized fetch enabled (ie: server a has content, blocks server b, server c boosts the content but does not block c, c boosts/upvotes/favorites it. A now doesn't know about any metrics from server C, despite being the original source of content).

while I'm all down for shaming the food industry into maybe not stuffing everything with added sugar, some of us need the amounts too.

As someone with type 1 diabetes, I need to know how much of everything (but especially carbs and kinds) to best guess how much insulin I need to dose before eating. Certainly there is a whole world of deceptive nutrition packaging to break down that doesn't help here, but the specific things you *must* report help to keep the labels honest enough to be workable.

I think the real issue is consumers have no idea what half of the ingredients are, or why they might or might not care about certain aspects of food. Nutritional science in general is very poorly researched. It's hard to get accurate data about what a diet does to people. Aside from locking someone in a room, feeding them everything they are to eat and controlling their activity (which paying volunteers in studies of that intensity is too costly for more than like n=1 studies), everything else is very unreliable. Many studies are backed up by surveys asking participants for what they ate for breakfast 6 months ago. As you might expect, no one remembers this crap, much less in useful detail.