Just for fun, I decided to try one method of converting a parabola to polar coordinates. Without loss of generality and assuming $a \neq 0,$ let's suppose our quadratic $y = ax^2 + bx + c$ can be factored as
$$y(x) = a(x-r_1)(x-r_2)$$
Now I notice that if our expression is of the form $u(u + c),$ it's easier to solve because we can then divide both sides. So, we make the substitution $u = x - r_1$ we end up with
$y(u + r_1) = au(u+r_1 - r_2),$ then if we convert to polar coordinates $y = r\sin(\theta),$ $u = r\cos(\theta)$ we can then divide both sides by $r \cos(\theta)$ to obtain
$\tan(\theta) = a(r \cos(\theta) + r_1 - r_2)$ which is then explicitly solvable for $r$ in terms of $\theta.$
But in going about this process, I'm confused as to how I "back-substitute" $u$ in terms of $x$ to complete the conversion process. Can $u$ substitution of this nature help simplify conversions to polar coordinates? Or is this concept flawed?