The coincident alignment of two galaxies on the sky can create the rare cosmic phenomenon of strong gravitational lensing, in which light from the more distant galaxy is bent around the foreground galaxy to create multiple, distorted, and magnified images. When the background galaxy hosts a bright active galactic nucleus, a quasar, the system becomes a probe of accretion disk physics, quasar-host galaxy relations, the Hubble constant, the stellar IMF, smooth matter fractions, amongst many other applications. It has been 40 years since the discovery of the first gravitationally lensed quasar, and dedicated spectroscopic and imaging surveys have added over one hundred new systems to this list. In recent years, the amount of available data across the whole-sky has grown exponentially. Full-sky data from X-ray to radio wavelengths exist, and predictions suggest there are many bright lensed quasars hidden in these datasets. This thesis presents several new techniques to mine these rare systems from whole-sky photometric datasets. We use the excellent resolving ability of Gaia, coupled with other wide-field surveys such as the Dark Energy Survey (DES), Pan-STARRS, and WISE, and present spectroscopic follow-up from the WHT, NTT, and Keck. By looking for multiple Gaia detections around photometric quasar candidates, and single Gaia detections near morphological galaxies, we have discovered 105 new lensed quasars. We also present a search based on significant offsets in astrometry and flux between Gaia and SDSS for spectroscopic quasars, suggesting several promising small-separation lens candidates. We characterise the confirmed systems based on ground-based imaging and the spatially resolved spectra, and comment on the purity, efficiency, and biases in our selection. DES data provides multi-epoch photometry over the baseline of years at optical wavelengths, allowing a colour-independent selection of lensed quasars by looking for nearby variable pairs. We create a parametric modelling pipeline of the DES images to extract lightcurves of system components, and show that it is a highly effective way to remove quasar and star projections before spectroscopic follow-up. We demonstrate that future searches based on detecting variability in multiple images will be biased towards four-image lensed quasars.