On the Suitability of Skyline Queries for Data Exploration

The skyline operator has been studied in database research for multi-criteria decision making. Until now the focus has been on the eciency or accuracy of single queries. In practice, however, users are increasingly confronted with unknown data collections, where precise query formulation proves dicult. Instead, users explore the data in a sequence of incrementally changing queries to the data to match their understanding of the data and task. In this work, we study the skyline operator as a tool in such exploratory querying both analytically and empirically. We show how its results evolve as users modify their queries, and suggest using our findings to guide users in formulating reasonable queries.