Methods, systems and articles of manufacture for discovering relationships among data elements within a dataset are disclosed. A first relationship is identified between a first data element and a second data element by identifying a correlation between a first attribute of the first data element and the first attribute of a second data element. A second relationship indicator is generated that is indicative of a relationship between the first data element and the second data element based on the correlation between the first attribute of the first and second data elements. Various embodiments can identify implicit relationships across one or more levels of explicit relationships where the explicit relationships can be across different attributes. Such techniques can be employed in various types of application programs.