Exploring Large Digital Library Collections Using a Map-Based Visualisation

Abstract

Abstract. In this paper we describe a novel approach for exploring large document collections using a map-based visualisation. We use hierarchically structured semantic concepts that are attached to the documents to create a visualisation of the semantic space that resembles a Google Map. The approach is novel in that we exploit the hierarchical structure to enable the approach to scale to large document collections and to create a map where the higher levels of spatial abstraction have semantic meaning. An informal evaluation is carried out to gather subjective feedback from users. Overall results are positive with users finding the visualisation enticing and easy to use.

Physical Metadata Visualisation: The Knitted Personal Library

“…Hanna Piotrowska [33] used Italo Calvino's conceptual book "If on a winter's night a traveller" to visualise themes (see Figure 2, right). A number of visualisation tools were developed to aid the exploration of collections [3,7,8,12,18,44] as well as exploration of in-document relationships and geographies [13,31,32,36] and support for playful exploration of digital collections [42,45]. Most of these tools remain bound to digital representations, and are therefore largely out of scope here.…”

Information and knowledge organisation in digital humanities

“…For example, Descy (2009) describes the use of map interfaces to visualise search result clusters. Similarly, Hall and Clough (2013) present an interactive map visualisation that enables the exploration of a hierarchical structure used to organise a collection of about 500,000 items taken from Europeana. In that interface, the elements on the map no longer represent real-world geography, but instead a virtual geography, where the map elements are concepts from the hierarchy.…”

Evaluation

“…Visualizations: the application included a novel approach for exploring document collections using a map-based visualization [16]. Evaluating visualizations is inherently complex and can include aspects such as the user experience, visual data analysis and reasoning, collaborative data analysis or work practices [22].…”

“…Subject hierarchies: to aid browsing and navigation of the collection the PATHS system included a subject hierarchy [16]. Data-driven approaches were developed to induce hierarchical topic structures automatically and enable the mapping of items from the collection onto the hierarchy.…”