Consumers flow is the life of a business district, and the stable and sustainable consumers flow is the key to the survival and development of a business district. How to find loyal and potential consumers in a business district, grasp their characteristics and carry out precision marketing is a difficult problem facing the survival and development of the business district. In response to this problem, communication operators can collect all user information in visiting business district through signaling data, and build a multi-dimensional user portrait system that includes customer portraits, social relationship portraits, and consumer preference portraits to grasp the characteristics of consumers, and according to Loyalty calculates loyal consumers, uses logistic regression to learn the profile characteristics of loyal consumers in different business district, and then uses random forest algorithm to identify potential consumers facing designated business district. After testing, the identification accuracy of the consumer profile construction and potential consumer mining method designed in this article with the business district as the user reaches 94%, which can accurately help each business district identify its own loyal consumers and potential consumers