I built a document recommendation project as part of a course, wrote it in python using the term-frequency inverse document frequency (TF-IDF) formula. It's actually a pretty straight forward method for recommending similar documents based on content.