This was a exam question so I know it cannot take too long to write out the proof. Only I cannot see an answer.
I would imagine you write $\sin(z) = \sin(1+(z-1)) = \sin(1)\cos(z-1) + \sin(z-1)\cos(1)$ and then use the everywhere-defined Taylor series for $\sin$ and $\cos$ to write $\frac{1}{\sin(z)}$ as the reciprocal of a power series. Then you manipulate it into the form $\displaystyle \frac{1}{1-P(z)}$ where $P$ is a power series and then invert using the geometric series formula. Only my $P$ looks horrible and thus the condition for convergence $|P(z)|<1$ is impossible to compute.
Another series which I cannot do but which I imagine could be done by similar methods is $\displaystyle \frac{1}{2\cos(z) -1}$ about $z_0 = 0$.
Any tips?
Using the Taylor series expansion for $\csc(x)=\frac{1}{\sin(x)}$, we may write write $$\frac{1}{\sin(z)}=\csc(z)=\sum_{k=0}^{\infty}\csc^{(k)}(1)\frac{(z-1)^{k}}{k!}.$$ This series will have radius of convergence $1$ since the poles of $\csc(x)$ are precisely at the the zeros of $\sin(x)$, all of which occur at integer multiples of $\pi$.
Now, this might not be satisfying at all, but this is likely the best that you can do. Ideally we would want to specify the coefficients $\csc^{(k)}(1)$ exactly (by giving a power series expansion we are doing precisely that) however they are extremely messy. The only way I can think of writing them without derivatives involves the Bernoulli numbers and powers of $\sin$ and $\cos$ evaluated at $1$.