It would probably most beneficial to you to use some ball and stick model to build the compounds you are interested in and to get some intuition about possible shapes. Once you know the shape atom distances for a given kind of a bond are to a good approximation constant, so finding possible dimensions is just a matter of reasonably simple geometry.
Depending on what you need you can try to estimate the molecule size from known shape and bond lengths (although you need to add a bit, as the distance between atoms in adjacent molecules is always larger than the bond lenghts), you can also try to estimate molecule volume from teh density (through molar mass and Avogadro's number). 2x4x8 Å is definitely in the right ballpark, but whether it is accurate enough depends on the application.
Angle part is a bit more difficult to address, as the molecule is not rigid, some bonds can freely rotate. That's where the ball and stick models come handy to help understand what is going on.