My early research tackled fundamental issues of databases such as data models, query languages, integrity constraints and transaction management dealing with deductive, object-oriented, logical-object, object-relational and active databases. This research led me to work on the other side of the data structure spectrum, i.e., video and textual data and to the integration of structured, semi-structured and unstructured data models and query languages. I then turned my attention to imprecise data management dealing with uncertainty and probabilistic data management. The results of this research were subsequently applied to analysis of medical and web data. More recently I have been studying social network analysis and predictive models for socio medical data. The former uses data analysis techniques which can deal with imprecise (structured, semistructured and unstructured) data, exploiting for instance the notion of conversation retrieval instead of the classic information retrieval techniques and new social network metrics for data propagation in multilayer networks. The later uses socio medical data for predicting (physical and cognitive decline) frailty in elderly. As side ef- fect we have also investigated supporting medical textual diagnosis using ICD-9-CM standard. The general goal of my research activity is to develop new techniques, driven by new application domains, to deal with imprecise data intensive applications. This has lead to create the SmartData Research Group within the University of Bologna.