Graphene is important in the study of 2D systems and has a number of unique properties and advantages: High charge-carrier mobilities and ballistic transport at room temperature, high structural stability, relativistic properties and a relatively simple production method. The potential of a tunable band-gap in graphene nanoribbons suggests that it could become a leading electrical component. One method that has emerged for modelling nanogaphene systems is the extended tight-binding model with Hubbard-\emph{U}. Within a real-space formalism, this model can be easily and efficiently applied to increasingly more complicated systems, where any number of edge defects, impurities and even patterning can be included, giving a more realistic description. This thesis investigated methods of structurally perturbing the ideal graphene nanoribbon device and probed the spin-dependent properties that arose: Random-edge vacancies, asymmetrical notches, uniaxial strain, magnetic inhomogeneity, chevron ZGNRs and patterned AGNRs. Random edge-vacancies have been used to perturb the electronic conductance in order to introduce the conductance gap observed in experimental results. These studies use the non-interacting tight-binding model, ignoring coulomb interactions. Introducing coulomb interactions within ideal ZGNRs has been shown to intrinsically include a conductance gap without edge-vacancies. The work presented in this thesis investigated the effects of edge-vacancies on the interacting model and demonstrated that, in general, the non-interacting model is insufficient to describe the physics of disordered ZGNRs. Controllable, asymmetric perturbations (i.e., notches and magnetic inhomogeneity) were added to interacting ideal ZGNRs to determine if the spin-dependent properties can be controlled. Asymmetrical perturbations exhibited spin-dependent conductance. In particular, magnetic inhomogeneity showed a transition from semi-conductive to half-metallic, suggesting a possible avenue for spin-filtering in spintronic devices. Finally, bottom-up synthesised GNRs were investigated (chevron ZGNRs and patterned AGNRs) and demonstrated controllable conductance properties and further work involving these systems was presented.