The Doppler shift in the light from the star tells you the period of the planet's orbit and also the velocity the star moves. You need to know the mass of the star, but this can be estimated to good accuracy from the star brightness and type. Once you know the mass of the star you can calculate the distance of the planet from it's period using:
$$ r^3 = \frac{GM}{4\pi^2}P^2 $$
where $M$ is the mass of the star and $P$ is the period of the oscillation.
Not that it's directly relevant to your question, but from the velocity of the star's oscillation we can calculate the minimum mass of the planet, because the velocity of the stars displacement depends on the gravitational force between the two. We can only calculate a minimum planet mass because if the plane of the system it tilted relative to us the true mass is higher than the one we calculate.
Having said this, these days most extrasolar planets are discovered because they transit their star, and these systems are not tilted relative to us (otherwise they wouldn't transit!). That means we can calculate an accurate mass for the planet.
In practice we normally turn the calculation over to a computer model (called a Bayesian Kepler periodogram if you want to Google it) because there are usually several planets and the oscillation is not a simple sine wave. We use a numerical fit to work out how many planets there are and how far from the star they are.