I create worlds and characters in my head more than I put on paper.  I put on paper more than I play.  I play far more than I DM.  My RPG experience is more of buying the books to read and then daydream about what I'd create.  For those times I go from what is in my head to actually writing it on paper, how much detail is needed where I feel I could leave it as is and move on to something else (because some day I'd like to DM in one of my created worlds)?  For items, should I create stats and work out the mechanics game-wise?  Is it okay to leave it open so I can have it do what I need it to do within my self-imposed guidelines, like the item is water based so I don't suddenly switch to fire?

For example, I've been working on a fantasy world where during a war 300 years prior a... weather vane like ballista was created.  Airborne creatures from other planes were a problem.  The device was to be autonomous, spinning and shooting at airborne creatures it considered hostile towards the continent (within parameters - it won't shoot a hawk, or raven, but air elementals and dragons are given).  It is powerful enough ancient dragons do not fly in the air space over this continent.  One shot kills them.  The ONLY time it shoots ground-based creatures is when the device itself is attacked.  It also shoots airborne creatures that may alter the future in a not so good manner.  It shoots a griffin that would harm a child that would be a future ruler.  As the future ruler, peace would be brought to the continent.  If the griffin wasn’t shot, the child never becomes a ruler and instead of peace there would be war.  That’s the general outline.  Then I freeze and wonder do I need to work it out more or move on to something else.  What are your thoughts on level of detail in a self-created world?  At what point do you say "Time to move on" or "I need to flesh this out more"?

One thing I do is always start with the adventure. Start with what the PCs will be doing, for the next few games anyway (you really can’t or even shouldn’t plan further than that), then flesh that out with world building. In the past, I’ve focused too much on world building and created some pretty poor adventures as a result. Starting with the actual game part makes sure that that won’t happen. Adventure first, then town/dungeon and local NPCs, then regional map and local history. Stuff like cosmology I straight up ignore unless it’s important, because most players are bored by it, or even feel like their characters are less significant to the DM as a result.

That said, I also like to write, so my “telling the story I want to tell” doesn’t get in the way of my “telling a story focused on and defined by the player characters.” That might help too!

Short answer though: if the players won’t care, it’s time to move on.

Edit: don’t be discouraged. If you enjoy world building on its own, that’s a hobby for some people! It doesn’t have to tie in to a book or game, often it shouldn’t even. If it’s fun for you, that’s all that matters. I know I enjoy just reading RPG sourcebooks once in a while...

I often think the best DMs are the ones who enjoy creating a world themselves as much as for the players. I would say do as much as you enjoy doing - make the maps and the stat blocks for things that energise you, and have a general idea at least if you aren't as enthusiastic about some part of it. I've found that having a story for the world in which the games are set helps open things up - for example, if the players decide to go off in a random direction instead of one of the ones I thought they would, or follow up a very left-field lead, knowing the greater story and history of the world means you can keep creating the adventure together - it's not derailing, it's exploring! Even if the players never delve into the history that you have in your head for the world, it's a really strong basis that you can build the adventure from on the fly. 

For items or creatures, if you know the narrative of how they are meant to live/work, then creating stats etc can come pretty easily if the party come across them, or you can quickly look up something similar enough that has a stat block. It sounds like you have a good idea of the mechanics, so if you don't want to create stat blocks etc for everything it's probably not necessary. I'd bet that having built a rich background and history, the abilities and stats would flow fairly naturally if you needed it. That said, you probably want a decent understanding of the rules so if you choose to ignore or bend them you do it consciously, not by mistake. That can give you more confidence as DM if you get a player pushing back against something you've created.

The only time you need to work out the actual stats is if the player characters will be expected to encounter it in game, and potentially interact with it in such a way that the stats are needed. For example, if you are just using the weather vane to explain why there are no dragons or air elementals on your continent to your players, and they will never see or interact with the weather vane, you don't need to stat it out. It can just be a "powerful item." If the weather-vane is non-functional after 300 years, and dragons are coming back, and the whole story of the campaign is to bring it back online so that it can save the continent from dragons, and you are going to end the campaign with a narrative after it is turned on that the dragons are all shot down from the sky and the good guys win, you don't need to stat it out. In these cases the weather vane is lore and a narrative tool, but doesn't need stats.

If you are going to have the player characters riding dragons and unicorns to invade the continent and take down the weather vane, now you need stats. You need to know what the saving throw is to avoid the hit, or what its to-hit bonus is against the dragons' and unicorns' AC. You need to know on what initiative count it will act. You need to know what its AC and HP are. You need to know the exact effects of this blast it is shooting at them (does it do a lot of damage? Or just Plane Shift the dragon into hell? Or what?) But you only need that if the PCs are going to directly interact with it in some way that requires the actual stats to be known. Otherwise, you don't need stats for it.

So my rule is -- if it could be interacted with during a game session in such a way that requires rolls to be made and actual effects to be known (usually, if it could realistically get into combat with the PCs, or be rolled against by the PCs, or roll against them itself), then the thing needs stats. If it's just part of the narrative landscape, it does not.

For example, my players are adventuring in a fantasy version of the Roman Empire and based in a Roman town. The town has walls, made of stone. And an iron-bound wooden gate. I did not stat out the walls and the gate. There is no real chance the players would ever attack them. And if a "war" happens with an army invading, I will not be rolling hp monster by monster to attack the town. I will use some more abstracted "army combat rules" to resolve unit vs unit combat. The walls would get some sort of a stat then, but it won't be AC and hp because I don't do unit combat that way. And I won't stat them out unless I need to.