The document discusses how mobility and fare purchase data from London's transit system could be used to build recommender systems to help travelers choose fare options that save them money. It finds that simple algorithms like naive Bayes and k-nearest neighbors can predict the optimal fare over 75% of the time, while decision trees achieve close to 98% accuracy. Currently travelers may overspend by £200 million by not choosing the best fare for their travel patterns. The data could help match travelers to the fares that are most cost effective for their individual mobility needs.