Discrete Dislocation Dynamics (DDD) simulations are accurate methods to calculate the plastic deformation of the materials at the micro-scale as they directly compute the dislocation interactions and their motion which are the underlying physical mechanisms of the plastic deformation of crystalline materials. DDD simulations are complex and data-intensive by nature, hence to interpret the material behavior through DDD simulations several post-processing and data analysis steps are required. Data mining and machine learning of dislocation simulations can play an important role in analyzing the large amount of data produced by those simulations. For instance, it is possible to gain insights on the effects of the dislocation mechanisms, which are not accessible through experiments, such as cross-slip mechanism. Also, surrogate models and hybrid simulations can remedy the high computational cost of these simulations by way of revealing the microstructure-property relationships and utilizing them in the solution of forward problems. In our study, we analyze the structure-property correlations in the microstructures by data mining and machine learning methods and investigate the important descriptors that define the behaviors and evolved microstructures in dislocation simulations. We also show our first promising steps towards replacing (parts of) DDD simulations by physics-based Deep Learning approaches.