I just self-learned the basics of using a Jacobian transformation on a double integral of a region to make solving it much simpler. I gave myself a sample problem I pulled arbitrarily, and I'm immediately very stuck.
My understanding is at least I should let $u = x + y$, because $x + y$ ranges from 2 to 5 in this region (x + y = 2 occurring at the intersection of the 2 other functions). Regarding whether or not this first step is even correct, or what other relationship I should use to let my other variable equal something, I'm very very lost now. Can someone help? Is it even possible to use a Jacobian for this double integral? Thank you for your time. (Note: I understand there are more roundabout ways of solving this with single-variable calculus, but I intend to understand Jacobians, not just solve this problem. Thank you)

You could take $u=x+y$ and $v=y-x$, which up to a scaling factor just means that you're rotating the coordinate system by 45 degrees, so that the Jacobian determinant will just be a constant. Then the boundary curve $y=x^2$ becomes $\frac{u+v}{2} = \left( \frac{u-v}{2} \right)^2$, which you can rewrite on the form $v=f(u)$ by solving a quadratic equation (choosing the correct sign in front of the square root), and similarly for the other curve $x=y^2$.