"Given a 15V input and desired outputs of 10V, 5V and 0V, how would I calculate the necessary resistance to use?"
I think a good way to go about this is to look at one set at a time. The standard voltage divider equation is simple enough, $$\text{Voltage across resistor of interest} = \frac{(\text{Resistor of Interest})}{(\text{Resistor of Interest + Resistor Not of Interest})} * V_{input}$$
When there are multiple nodes, like in the example you've given, just simplify it to the basic resistor divider and find the first voltage. Alternatively, if we're given voltages, we can rearrange this equation to solve for the resistor of interest in terms of the resistor not of interest.
$$\text{Resistor of Interest} = \frac{1}{({V_{input}}\div{\text{Voltage across resistor of interest}})-1}*\text{Resistor Not of Interest}$$
To simplify, in your example for the 10V node, the resistor of interest is the combination of R2 and R3, leaving the resistor not of interest as R1. Once you've found your ratio between (R2+R3) and R1 you can move on to find the ratio for R2 and R3. In this case you can just look at those two as another divider and the input voltage is that first node voltage you've just used as your output voltage. Following this method you'll find that R1 is one third (R2+R3) and that R2 is the same as R3. It makes sense that given equal current flow, an identical drop across each resistor means and identical resistance, following Ohm's law V=IR.
"Is it possible to create a voltage divder that does not have proportional drops (e.g., let's say that from this same circuit, I want 14V, 12V, 5V and 0V)?"
This will be the same process as before, but just plug in different voltages. For the first node:
$$\text{(R2+R3)} = (\frac{1}{(14V\div12V)-1})*\text{R1}=6*R1$$
So the combination of R2 and R3 is six times larger than R1 alone. For the second node:
$$\text{(R2)} = (\frac{1}{(12V\div5V)-1})*\text{R3}=0.71*R3$$
Finally, and this is the most difficult part for most students, just pick a resistor value. This is the engineering part of electrical engineering, you have to make a decision. This one is not too difficult, for the most part larger resistances are better. Larger resistances will reduce current flow while still providing the voltages you need.
There are several other considerations when using a voltage divider in practice. These are great for basic reference voltages or proportionally knocking down a signal voltage in a single direction. For instance a 5V signal being taken down to 3.3V for a microcontroller works well because a voltage divider acts like an attenuation coefficient to the signal, everything gets reduced by the same amount.
If you're proving voltage to a device of some sort, you can sometimes model that current draw as a resistance, assuming it's always constant (R=V/I). This device resistor, or load, is usually the resistor of interest or parallel to the resistor of interest. I would not recommend this at any time however as the node voltage will change depending on the current draw of the load.
"And how does that math work?"
See equations above.