Using Google Earth Engine and documentation in this linked repository, I developed a classification system for the whole Amazon rainforest, validated using Hansen's Global Forest Change dataset.
This project was developed for the final of one of UNC's Special Topic Computer Science courses: COMP 590-800 Data Science for Earth Sciences. In the JS file, I use earth engine capabilities and supervised machine learning algorithms to delineate training regions, analyze those regions' satellite spectral data, and classify the change in these testing regions as forest, forest loss, forest gain, or non-forest.
Additionally, attached is my final paper, detailing the problem and the applications that this research can accomplish.