Visualization in information retrieval: a three-level analysis

Abstract

Visualization in information retrieval (IR) is a rapidly emerging area in which human-computer interaction (HCI)is adapted to the design of interfaces for IR systems. There are numerous proposals for the development of new visual IR systems. In parallel with the advent of advanced software and hardware, the attempts to design visualization-based IR systems are stimulated with the promise that information visualization applies visual processing to abstract information. Under this rapid development in the research area, there are few attempts to conceptualize this new field of study. This paper intends to classify the works of visualization for IR system design in a systematic way, by providing a mosaic view of the emerging area of visualization in IR system design. The paper proposes a three-level approach to analyzing visualization in IR system design: (i) component level analysis, (ii) technique level analysis and (iii) interaction level analysis. This three-level approach helps us to understand the state of visualization for the design of IR systems in multiple ways. The three-level approach also enables us in further serious investigation in this emerging field that has been created by the intersection of visualization and IR, by allowing us to look at different layers of visualization in IR that have both technical and theoretical concerns. 50 1 2

Exploring the context of visual information seeking

“…However, as information is rapidly growing in quantity, heterogeneity and dimensionality, designing such systems has become an increasingly difficult task (Gerken et al, 2009). As the interaction between users and visual representations takes place during information seeking, a poor representation could not only burden users more but also disturb their information-seeking process (Song, 2000).…”

A framework for evaluating geographical information

“…Historically, cartographers have designed maps to fulfil this function; in a digital environment geovizualization has been developed to promote the exploration of GI [35]. This activity is developing in parallel with work on visualization in information retrieval [36].…”

Progress in visualisation

“…Song [11], in an article with 120 references, shows that there is no shortage of activity in this field. If visualisation is to become widely adopted for information retrieval, as Song says, '[it] must be preceded by both knowledge of textual analysis and data conversion and it has been identified that the decision in choosing a visualization technique must be made on the basis of understanding data types and data representations.…”