Availability and frequency of synthetic aperture radar (SAR) imagery are rapidly increasing. This surge of data presents new opportunities to explore both old and new questions relating to surface deformation. This expansion also introduces common challenges associated with large volumes of data. I present a database of earthquake slip distributions derived from geodetic observation. I extract common spatial descriptors from these slip distributions and regress these spatial descriptors with moment magnitude to derive new scaling relationships. I find that my scaling relationships differ in important ways from previous studies and show that these differences originate from my use of a geodetic slip distribution database rather than from methods for extracting spatial descriptors. I present a convolutional neural network trained to detect, locate, and classify the presence of co-seismic-like surface deformation in interferograms. The results show that the network obtains an overall accuracy of 99.74% on synthetic images, and 85.22% on real images. I present a convolutional denoising autoencoder trained to reduce noise in InSAR interferograms. I show that this network has the capabilities to learn features of surface deformation within InSAR interferograms and reduce the noise in the interferogram with little loss of signal.