Graphene nanoribbons (GNRs) - one-dimensional strips of graphene - share many of the exciting properties of graphene, such as ballistic transport over micron dimensions, strength and flexibility, but more importantly, they exhibit a tunable band gap that depends on the atomic structure. Recent advances of fabricating atomically precise bottom-up GNRs with unprecedented control over their atomic structure have attracted interest from the field of nanoelectronics. However, a big part of the future success of GNRs depends on the ability to produce and integrate GNR junctions into complex next-generation devices. Much more effort is needed in both perfecting the production techniques and improving the theoretical understanding of these exciting nanostructures. This thesis is, therefore, devoted to exploring electronic transport properties in graphene nanoribbon junctions and unraveling their underlying structure-property relationships. Using tight-binding models, density functional theory and Green's function method, we determine the electronic properties of both experimentally synthesized and theoretically proposed junctions. In the first part of the thesis, we examine width-modulated GNR nanostructures and discover a subtle interplay between the localized states in the scattering region and the continuum of states in the leads. We show that depending on the size of the scattering region, we observe contrasting behavior on the electronic transport properties. Next, we expand the width-modulated region and show that a width-dependent transport gap opens in the presence of a quantum dot, thereby yielding built-in one-dimensional metal-semiconductor-metal junction. Part II of this thesis is dedicated to a joint experimental and theoretical effort in order to reveal the detrimental effect of ``bite'' defects, resulting upon the cleavage of phenyl groups of precursor molecules. We explore their effect on the electronic transport from first-principles calculations and show how conduction is disrupted at the band edges. We then generalize our theoretical findings to other nanoribbons in a systematic manner, thus establishing guidelines to minimize the detrimental role of such defects. Later, we show that strategically placed ``bite'' defects can selectively modify electronic transport properties and apply this concept to construct two prototypical components for nanoelectronics. Whereas, in Part III we employ high-throughput screening of over 400000 angled junctions in order to find potential candidates for interconnects in logic circuits and determine design rules based on structure-property relationships. We discover that the bipartite symmetry of graphene lattice and the presence of resonant states, localized at the junction, play an important role in determining the transport properties of angled junctions. Besides, we also provide a web application that allows easy design and calculation of electronic properties of GNR junctions. Finally, the last chapter of the thesis involves developing a more realistic model for transport calculations by including finite length and contact effects in order to reduce the gap between the experimental and theoretical results.