Illustrations like this one only serve to muddy the waters - http://witchy.co/triangle.png

The moon is lit from the wrong side and it has a crescent shape in what's basically an overhead projection. And it's not like there's much else going on in the illustration, it's one of its main elements.

You're not wrong, and for the amount of effort put into this presentation, I'm surprised at this oversight. But if you can think of this image as a schematic, it's not a huge problem.

Some time ago I calculated moon phase for date of birth for 90k users and I found no correlation.

I calculated gpa vs. date of birth, there is a very slight skew towards post-july being more apt in terms of gpa.

Not sure if significant.

That would imply a student being towards the upper end of the age cut off for starting kindergarten/elementary school (in the US at least, varying by state).

If you calculated the moon phase from the date of birth how is it possible that the two are not correlated?

I didn't found higher birth rates for any of the phase (split into quarters)

9K of Javascript to handle common Sun, Moon calcs: https://github.com/mourner/suncalc

Fwiw it's 9KB, not 9K SLOC.

Is this a common misconception? Personally, I assume KB in the context given.

The initial diagram showing new/half/full moon goes off my screen, presumably because I'm at 150% desktop zoom and it possibly somehow gets confused by that?

Also the feedback about the wrong side being illuminated.

Very hard to read on my iphone

The font seems to have an extremely strong glare effect in combination with other colors on the page.

The font is made of sublimated astral magick.

Very nicely done.