In a couple of ongoing and newer projects within agriculture as well as social georgraphy we are exploring satellite data from Sentinel 1 and Sentinel 2 in order to ascertain ground conditions. In both of these projects clouds and other obstructions impeed our machine learning algorithms to learn. We need to process thousands of current image band satellite data through convolutional type of neural networks and remove obstructions (such as clouds) based on relevant temporal and spatial information. The end result would be an advanced type of interpolation where the resulting image would have the closest possible cloud-free information. Using this resource has proven instrumental in our work, as can be seen from its extensive use during the last year, and we now wish to continue using it at a much higher capacity (i.e. more memory).