You could take a look at Srednicki's book (draft copy free online, google to find it), which starts with reps of the Lorentz group and slowly builds up to the Dirac lagrangian.
Incidentally, the Dirac equation can't really be derived, it is just postulated as following from the simplest lagrangian (that is, terms with the fewest derivatives) that can be written down for a field corresponding to spin-1/2 particles.
For the extreme version of this point of view, see Weinberg's book (which is thorough and extremely detailed, and therefore comprehensible, but only with a lot of effort).