Using dialog and context in a speech-based interface for an information visualization environment

Abstract

We describe a speech-based interface to an information visualization (infoVis) system. Users ask natural-language questions about a given data domain. Our interface then maps the questions into infoVis operations, which result in the display of data visualizations that address the questions. Users can interact with these views via speech or direct manipulation. If users give incomplete information, our interface guides them in clarifying their questions. The intelligence behind our interface is encapsulated in a service logic that embodies domain knowledge about both the data being explored and the infoVis system. This allows users to focus on answering questions, rather than on the mechanics of accessing data and creating views.