Aspect grid: A visualization for iteratively refining aspect-based queries on document collections

Abstract

Search results comprising of unstructured documents are traditionally displayed as plain lists, whereas the formulation of complex queries can be aided by comprehensive visualizations that highlight relevant attributes of the search results. We propose a visualization concept that incorporates a sentiment analysis of aspects mentioned in examined documents. Our concept provides a means of iteratively refining the query based upon preliminary search results. After an abstract description of the concept, we elaborate on its usage in a practical example of filtering customer reviews in order to choose a restaurant based on user-defined preferences.