Toward interactive visualization of results from domain-specific text analytics

Abstract

In big data analytics, visualization and access are central for the creation of knowledge and value from data. Interactive visualizations of analysis of structured data are commonplace. In this paper, information visualization and interaction for text analysis are addressed. The paper motivates this issue from a data usage standpoint, gives a survey of approaches in the area of interactive visualization of text analytics, and presents our proposal of a specific solution design for visual interaction with results from a combination of named entity recognition (NER) and text categorization (TC). This matrix-based model illustrates abstract views on complex relationships between abstract entities and is exemplary for any combination of feature extraction and TC. The aim of our proposal is to support feature extraction and TC researchers in distributed virtual research environments by providing intuitive visual interfaces.