If you aren't completely comfortable working with quotient spaces, the following is very useful.
Universal property of the quotient space
Let $X,Y$ be topological spaces, and let $\sim$ be an equivalence relation on the points of $X$. Then we have a natural continuous map $p\colon X\to X/{\sim}$ given by projecting a point of $X$ on to its equivalence class.
Now, let $g\colon X/{\sim}\to Y$ be a continuous map. By composing with $p$, we get a continuous map from $X$ to $Y$:
$$
g\circ p\colon X\xrightarrow{p}X/{\sim}\xrightarrow{g}Y
$$
Write $f=g\circ p$. Now $f$ isn't just any old continuous map from $X$ to $Y$; it has a special property; namely, that if $x\sim y$ then $f(x)=f(y)$. [This is because $p(x)=p(y)$.]
Why is this interesting? Well, it gives us a characterization of the continuous maps from $X/{\sim}\to Y$:
if $f\colon X\to Y$ is a continuous map such that
(*) $f(x)=f(y)$ whenever $x\sim y$
then there is a unique continuous map $g\colon X/{\sim}\to Y$ such that $f=g\circ p$.
In other words, the following map of sets is a bijection:
\begin{align}
\{\textrm{Continuous maps }g\textrm{ from }X/{\sim}\textrm{ to }Y\}&\to\{\textrm{Continuous maps }f\textrm{ from }X\textrm{ to }Y\textrm{ satisfying (*)}\} \\
g&\mapsto g\circ p
\end{align}
So, coming up with a continuous map from $X/{\sim}$ to $Y$ is the same thing as coming up with a continuous map from $X$ to $Y$ that is constant on equivalence classes.
This particular example
Here, you want to define a homeomorphism from $\mathbb R^2/{\sim}$ to $S^2$, where $\sim$ identifies all points whose norm is at least $1$. By the above discussion, we need to specify a continuous map from $\mathbb R^2$ to $S^2$ that is constant outside the open unit disc. Clearly, in order to do this, it is enough to define a continuous map from the closed disc $D^1=\{P\in\mathbb R^2\;\colon\;\|P\|\le 1\}$ to $S^2$ and then extend it by the constant value everywhere else.
Can you think of a suitable map? It should be surjective, and it should be injective except on the boundary of the disc, where it should take some constant value.
Hint: Remember that the closed disc $D^1$ can be parameterized by polar coordinates $(\rho,\vartheta)$, and that the sphere $S^2$ can be parameterized by spherical polar coordinates $(\phi, \theta)$.
\|rather than||for the norm; and use{\sim}to remove the excessive spacing around\sim. – Asaf Karagila Aug 01 '15 at 14:26