I have somewhat strong opinions on this. I uploaded a similar set of symbols here:
https://imgur.com/a/gKJ3KO2I’ve added notes in red, some baselines in blue and some lines at x-height in yellow.
I want to point out a few things:
1. OP misses out the need for script and blackboard bold letters and various symbols. You want them to not be confused with letters. In particular U and set union, epsilon and set membership, and oplus and capital theta.
I think OP’s eta looks wrong. Too much like an m.
One trick to use is that letters can have ascenders and descenders. Numbers can too but I don’t use that. I also incorrectly write a rho as a letter with an ascender and no descender to help to distinguish it from a p. I also try to impart a big curve at the top so you won’t think there could be the top of a straight line hidden in there.
I draw the top of a tau just below x-height while the bar on a t is somewhat above it. This and the rest of the t being higher helps to disambiguate.
Never use an o (“oh”) or omicron and probably not upsilon either (I missed out capital upsilon because OP did too and I didn’t notice). An exception is big/little O notation but I don’t really like it.
I don’t like OP’s q. It looks too much like a double bowled g. I prefer a tick to a loop.
I tend to not have issues with 2 vs z but I do with nu vs v. I kinda dislike using nu because of it. In print I have trouble with v vs u.
I often use variant letters for some Greek letters (in particular my epsilon is what TEX calls varepsilon because it is more like handwriting and less like a print font. Similarly for theta kappa, and phi. I don’t use varpi which looks too much like an omega, cardigan which is only really to go on the end of a word in Greek, varrho because I don’t like it, and I don’t write capital omega the modern Greek way (a horizontal line with a circle above it).
Other tricky symbols are:
Angle brackets vs lt/gr. Usually context suffices. Angle brackets vs parens: just don’t rely on this distinction.
Or vs v can and union vs U can usually be solved by context and spacing (you have a bigger space between a term and an operator than between factors or arguments in a term)
A final note is that I tend to write mathematics more upright than my normal handwriting. This helps distinguish inline mathematics from the rest of a sentence.
I looked through some random old notes to see if I could still read my writing and I can. The only thing I struggled with was the script C for conjugate classes because I was missing the context and struggling to guess what ccl stood for. At first I thought it was a script G.
Having written this, I realise I omitted forall, exists, nabla/del, integral sign, partial sign, therefore dots, infinity, approxequal and hbar.
A final note is that spoken names matter to o. “Twiddle” is a much better name than “tilde” because it is more naturally made into a verb. You can say “define the relation twiddle on the naturals such that a twiddles b if ...” and then you can easily use that verb when talking through a proof of eg transitivity.