Experts estimate that the world loses 48 football fields of forest every minute. The Amazon rainforest represents over half of the worlds remaining forests.

Planet, open-sourced roughly 54Gb of satellite imagery taken over the Amazon rainforest. The goal was simple: use the satellite data to track the human footprint in the Amazon rainforest.

This repostitory documents my use of TensorFlow to build a 3-layer convolutional neural network to classify satellite imagery.

For more information please reference: Understanding the Amazon from space