The best view is always from the top. This holds true for the constellation of satellites hurling millions of pixels in images down to earth every day. In this age of breathtaking photography advancements, from drones and action cameras to submersible mobile phones, what has changed for an almost 60-year-old technology? It turns out that deep learning, i.e. an advanced machine learning technique that develops multilayered mathematical structures to classify images, is changing the way computers interpret large sets of satellite data. DigitalGlobe, a Colorado-based satellite imagery provider that supplies to such clients as Google Maps, is using a cloud-based, geospatial big data (GBD) platform to intelligently extract relevant images and points of interest from a vast compendium of above-earth photos.