Database queries can return voluminous answers that are time consuming to read and that provide information with too much detail. A solution to this problem is presented in this paper. The solution utilizes approximate domain abstraction to construct a hierarchy of databases. Each successive database utilizes a more abstract domain language than the previous one. Each successive database is also an approximation of the previous one. Users may interact with the system to gradually refine concise, abstract, approximate answers into detailed, lengthy, exact ones. Users may also utilize the system to construct concise hierarchically structured answers to database queries. Implementation of this method for datalog databases is in progress. Expected results include empirical tests of the method on real world databases.