As he is at pains to point out in the video, you need to be careful about the meaning of the symbols. Sometimes $x$ and $y$ are functions, and sometimes co-ordinates, depending on context within the equations. He's talking about the transformation from cartesian co-ordinates $(x,y)$ to polar co-ordinates $(r,\theta)$, and the inverse transformation from polar co-ordinates $(r,\theta)$ back to cartesian co-ordinates $(x,y)$. The forward transformation is a function $P: \mathbb R^2 \to \mathbb R^2$ defined by
$P(x,y) = (r(x,y),\theta(x,y))$, where $r$ and $\theta$ are functions defined as
$$ r(x,y) = \sqrt{x^2 + y^2}$$
$$\theta(x,y) = arctan(y/x)$$
and the inverse transformation is defined as $P^{-1}(r,\theta) = (x(r,\theta),y(r,\theta))$, where here $x$ and $y$ are functions defined as
$$x(r,\theta) = r \cos \theta$$
$$y(r,\theta) = r \sin \theta$$
If we compose the forward and inverse transformations, we get the identity
$$P^{-1} \circ P(x,y) = (x,y)$$
or, as functions
$$x(r(x,y),\theta(x,y)) = x$$
$$y(r(x,y),\theta(x,y)) = y$$
In the video he goes on to calculate the Jacobeans of $P$, $P^{-1}$ and $P^{-1} \circ P$, which as these are all functions from $\mathbb R^2$ to $\mathbb R^2$ will be 2x2 matrices formed from the partial derivatives of the co-ordinate functions. Now differentiating this function $x$ with respect to the co-ordinate $x$, and using the chain rule, we get
$$\frac{\partial x}{\partial r} \frac{\partial r}{\partial x} + \frac{\partial x}{\partial \theta}\frac{\partial \theta}{\partial x}= 1$$
which is easily verified by working out each of the partial derivatives and expressing everything in terms of $r$ and $\theta$.