The data for this project comes from three different datasets. Twitter archives, containing roughly 2.5k tweets and other information; image predictions, which stores about 2,000 entries of image predictions; and tweet details, that were obtained quering the Twitter API.