Exploratory analysis of large web datasets

Abstract

In the era of big data, the capability to identify very quickly prominent summary information about a target entity of interest, like a person or an event, from large datasets is essential, and exploratory analysis techniques help in this direction. In this paper, we provide a solution based on smart entity views and on pre-defined analysis operators which exploit keywords available in the entity view together with similarity information to produce summary information about the view contents from both a thematic and analytics perspective. In particular, smart entity views can be analyzed according to the following exploratory paradigms: entity expansion, entity visualization, and entity analytics. The proposed approach is discussed by referring to a case study of twitter dataset related to the “Expo2015” event as target entity.