It uses t-SNE but there other other working parts here, including word2vec (and some nice compression of a pre-trained model), keyword searching to provide cont

It uses t-SNE but there other other working parts here, including word2vec (and some nice compression of a pre-trained model), keyword searching to provide context for terms, and clustering to find natural dense groups. It's a nice pipeline that fits together a bunch of independently interesting parts into a single system that can produce quite remarkable results.