Advances in information technology prompt a tremendous usage growth of the Internet. Online activities, such as e-commerce, social interaction, etc., have drawn increasing attentions in regard to the provision of personalized services which require best and comprehensive understanding of users. As an approach, this study outlines a general framework based on human (or consumer) contexts for the discovery and creation of business intelligence. Three major portions are discussed. First, the collection of human contexts, including activity logs in both cyber and physical worlds, is modeled. Second, data analysis was performed via proposed mining algorithms that concern potential fusion at different levels according to situations and ultimate purposes. Third, sustenance of developed model is then concentrated. An open platform was developed to support the evolutionary process of human models, and to allow contributions (e.g., data sharing, accessing, etc.) from third parties.