When combined with land price data, with environmental data, with GDP growth data, and all sort of data, people can train a machine learning algorithm together with all the data controllers to predict how likely some company’s engaging in shell company behavior for the next quarter.