Very cool!
ok, so here are some things. I have been doing the same thing you have been doing, but I have been doing it for about 2 almost 3 years now, so I'll give you some heads up.
1st, the uncanny valley, if you have not heard of it before, it is referring to the effect of how robots, as they improve, start to look more and more like humans, their is a point when robots look almost human, but are not quite there yet. Infact, they start to look disturbing. The idea is that, they look human enough to trick our brains, but our brains are also program to identify anything that looks wrong incase there may be something wrong with the person we are looking at. Here is how I think the property can be applied to art. The more realistic you try to be, the less room you have for error, the easier it is for your work to look bad. This leads to another issue, often, to fix the issue, you basically have to correct something in a manner that involves making the mistake look realistic. The issue with this is now your work is even more realistic, so you have even less room for error, then something else will look off, so new issues show up that weren't an issue before. its like making your bed and puting on a new sheet, but every time you slip the last corner on, another comes off. So what my point is, you will need to at some point say to yourself, this is good enough, you will need to set some things to be restrictions on what you will and will not change or try to improve on if you ever want to complete something.
2, not sure if this will really help you since you can just ignore it if you want, but head ratio. If you haven't heard about it yet, just do a quick google image search. Now I wouldn't say copy what you see in those images, since playing around with proportion ratios is always good fun and interesting, but one thing you may want to play with is the character's final height. from what I can tell, over the past 8 or 12 year, the anime and cartoon industry, as well as the art community seems to have settle on a 7.5 head height as the ideal ratio for how tall a drawn character should look, and it works for 3D as well. note this is for the idealized body, not a realistic body. Ofter, however, this rule needs to be broken, to show off a characters age or just exaggerate their body type.
3 it looks like you have an understanding of color theory a bit, at least with the shading, so I won't comment on that, but I will mention, it looks like you are using something like an emittion node with colors being controlled by maybe a normal vector node or/and a frensel layer weight node. I too have tried this, and I think it is fine for the most part, I still use them to add effects to characters or the enhance them beyond a bland texture, but I never did settle for this style. I don't know why, if it didn't meet my expectations, if it couldn't adapt to lighting well, or if I just couldn't commit. I am currently working on emulating the disney style, and I fine that with good lighting techniques, while it takes more time, can achieve the same thing, if not better. that is, if you like the disney still, your style still works better if you want to capture that anime feeling. Oh, now I remember why I didn't stick with it. Because in drawn art, to make a point clear, you have to simplify and exaggerate, and that even includes lighting, that means adding shadows where they shouldn't show up, and light where there shouldn't be light, and often, some effects look good but were just the results of an artist cutting corners in order to push out enough frames to get their quota for that day. There is a lot of junk when it comes to how anime is lit and 2D drawn techniques, and the mathmatical perfection of 3D kinda mess that up, which is why I gave up on the technique you are using. idk, guess I was just going to warn you about it. Oh no, never mind, I was trying to give you the advice that, if you should play with lights, learn how to adjust their size and the power of making lights harder or softer to replicate the animated effects you want.
4 if you want to make your art feel more 2D, there are some things to do and not to do, avoid front (isometric) and side (isometric) views. Not that isometric is bad, it can work with some models a lot of the time, (I even use perspective lenses as large as an 85mm for things like a face shot, but I keep 55mm as my low point unless I am in tight spots) but it seems that with front and side shots, you are giving users too much information about your models, enough that their brains can subconsciously criticize your works. I would say, for angle shots, if the idea is that the character is turned away from the camera, but you are still looking at them, it is fine to adjust the head so that it looks better, angled up to be level with the camera and turned and maybe even tilted to or away from the camera to indicate if we are looking at them, or if they are looking at something else. if you are going to move the eyes, I have given up on an ik rig for them, it often looks better if you adjust each to match how eyes are often drawn in anime (from different angles) because when it's realistic vs stylized, this is one of the cases where the eyes seem to make more sense to our weak brains when the eyes are moved to a more stylized position rather than a realistic one (which is another thing to try and balance and why this stuff is a pain to do).
if you actually managed to read that wall of barf, thanks perhaps it will save you a month or more of frustration.
oh, maybe one last thing, it seems like (for me) eyes always look better just to have a little bit of an outline on them (black), so that may something you want to play with (but does not work well with the anime style so I gave up on it, but you could try). Best of luck!
